Deep_Learning_Techniques.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "#Requirements"
      ],
      "metadata": {
        "id": "4EazC9brZDtR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install tensorflow==2.12.0\n",
        "!pip install scikit-image\n",
        "!pip install scikit-learn\n",
        "!pip install opencv-python-headless\n",
        "!pip install numpy\n",
        "!pip install matplotlib\n",
        "!pip install glob2\n",
        "!pip install pickle5\n",
        "!pip install google-colab"
      ],
      "metadata": {
        "id": "gcua68_XN1QQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vD7ifqACvYuL"
      },
      "source": [
        "#Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8-UqaUr4Y5MH"
      },
      "source": [
        "import numpy as np\n",
        "import cv2\n",
        "import matplotlib.pyplot as plt\n",
        "import glob\n",
        "import os\n",
        "import pickle\n",
        "import tensorflow\n",
        "from tensorflow import keras\n",
        "from keras import layers,models\n",
        "\n",
        "\n",
        "from skimage.io import imread, imshow\n",
        "from skimage.color import rgb2hsv, rgb2ycbcr\n",
        "from skimage.transform import resize\n",
        "\n",
        "from sklearn import metrics\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import classification_report, confusion_matrix, cohen_kappa_score, roc_auc_score\n",
        "\n",
        "\n",
        "import warnings\n",
        "def ignore_warn(*args, **kwargs):\n",
        "    pass\n",
        "warnings.warn = ignore_warn\n"
      ],
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "brv2ZoGQ2Vrv",
        "outputId": "060ffff7-f78c-45e7-a9d6-c7c048227e04"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q_ILJudQve9c"
      },
      "source": [
        "#Convolutional Neural Network (CNN)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Data preparation"
      ],
      "metadata": {
        "id": "Bp3KOdqdp2yY"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Extract images from the videos\n",
        "\n"
      ],
      "metadata": {
        "id": "ZUx3O4EpqGOQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#For potential data augmentation\n",
        "\"\"\"def rotation(image, seed=777,rota=0):\n",
        "    random.seed(seed)\n",
        "    angle= rota\n",
        "    r_img = rotate(image, angle, mode='reflect', reshape=False, order=0)\n",
        "    return r_img\n",
        "\n",
        "def v_transl(image, seed=777,decalage=0):\n",
        "    random.seed(seed)\n",
        "    n_pixels = decalage\n",
        "    vtranslated_img = np.roll(image, n_pixels, axis=0)\n",
        "    return vtranslated_img\n",
        "\n",
        "def h_transl(image, seed=777,decalage=0):\n",
        "    random.seed(seed)\n",
        "    n_pixels = decalage\n",
        "    htranslated_img = np.roll(image, n_pixels, axis=1)\n",
        "    return htranslated_img\"\"\""
      ],
      "metadata": {
        "id": "ukUDmr7J9Anp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nuEjrMNLayPy",
        "outputId": "a83f3c8e-768d-40d9-d3ba-fc29d445ffc7"
      },
      "source": [
        "label2int = {\n",
        "    'dislike' : 0,\n",
        "    'like' : 1,\n",
        "    'ok' : 2,\n",
        "    'point' : 3,\n",
        "    'slide_1' : 4,\n",
        "    'slide_2' : 5,\n",
        "    'take_a_picture' : 6\n",
        "}\n",
        "\n",
        "X_set = []\n",
        "Y_set = []\n",
        "for path_director in glob.glob('/content/drive/MyDrive/demo_room/*'):\n",
        "\n",
        "  label1 = path_director.split('/')[-1]\n",
        "  print(label1)\n",
        "  for image_path in glob.glob(path_director+'/*'):\n",
        "      cap=cv2.VideoCapture(image_path)\n",
        "\n",
        "      label = image_path.split('/')[-1].split('_')[0]\n",
        "      if label=='slide' or label=='take'  :\n",
        "        label=label+'_'+image_path.split('/')[-1].split('_')[1]\n",
        "        if label=='take_a':\n",
        "          label=label+'_'+image_path.split('/')[-1].split('_')[2]\n",
        "      nb=0\n",
        "      while cap.isOpened():\n",
        "\n",
        "        imageId=cap.get(1)\n",
        "        ret,image=cap.read()\n",
        "        if ret !=True:\n",
        "          break\n",
        "\n",
        "        if imageId != 0 and imageId % 4 == 0:\n",
        "            nb=nb+1\n",
        "            image=cv2.cvtColor(image , cv2.COLOR_BGR2RGB)\n",
        "\n",
        "            if image.ndim==3:\n",
        "              image = cv2.GaussianBlur(image, (5,5) , 100)\n",
        "              red , green , blue = image[:,:,0] ,image[:,:,1] , image[:,:,2]\n",
        "\n",
        "              #get the HSV version\n",
        "              image_hsv = rgb2hsv(image)\n",
        "              hue = image_hsv[:,:,0]\n",
        "              sat = image_hsv[:,:,1]\n",
        "\n",
        "              #get the ycbcr version\n",
        "              image_ycbcr = rgb2ycbcr(image)\n",
        "              y = image_ycbcr[:,:,0]\n",
        "              cb = image_ycbcr[:,:,1]\n",
        "              cr = image_ycbcr[:,:,2]\n",
        "\n",
        "              \"\"\"#HSV+RBGA mask for hand extraction\n",
        "              mask = (hue>=0) & (hue<=0.13888) & (sat>=0.23) & (sat<=0.68) & (red>95) & (green > 40) & (blue > 20) & (red > green) & (red > blue) & (np.abs(red - green) > 15)\n",
        "              \"\"\"\n",
        "              #YCBCR+RGBA mask for hand extraction\n",
        "              mask = (red > 95) & (green > 40) & (blue > 20) & (red > green) & (red > blue) & (np.abs(red - green) > 15) & (cr > 135) & (cb > 85) & (y > 80) & (cr <= (1.5862*cb)+20) & (cr >= (0.3448*cb)+76.2069) & (cr >= (-4.5652*cb)+234.5652) & (cr <= (-1.15 * cb)+301.75) & (cr <= (-2.2857*cb)+432.85)\n",
        "\n",
        "              mask = mask.astype(np.uint8)\n",
        "              mask[mask == 1] = 255\n",
        "              np.sum(mask == 255)\n",
        "\n",
        "\n",
        "              kernel = np.ones((5,5), np.uint8)\n",
        "              mask1 = cv2.erode(mask , kernel , iterations=2)\n",
        "              kernel = np.ones((5,5), np.uint8)\n",
        "              image = cv2.dilate(mask1 , kernel , iterations=4)\n",
        "              image=resize(image,(128,128))\n",
        "\n",
        "              X_set.append(image)\n",
        "              Y_set.append(label2int[label])\n",
        "print(len(X_set))\n",
        "print(len(Y_set))\n"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "subject_5\n",
            "subject_4\n",
            "subject_3\n",
            "subject_1\n",
            "subject_2\n",
            "6029\n",
            "6029\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qkud5mUDvtbM"
      },
      "source": [
        "###Save the obtained data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XoUz4jnmp293"
      },
      "source": [
        "X=np.array(X_set).reshape(-1,128,128,1)\n",
        "Y=np.array(Y_set)\n",
        "\n",
        "pickle_out=open(\"X_data_CNN.pickle\",\"wb\")\n",
        "pickle.dump(X,pickle_out)\n",
        "pickle_out.close()\n",
        "\n",
        "pickle_out=open(\"Y_data_CNN.pickle\",\"wb\")\n",
        "pickle.dump(Y,pickle_out)\n",
        "pickle_out.close()"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4uuv-Jtss3pC"
      },
      "source": [
        "pickle_in1=open(\"X_data_CNN.pickle\",\"rb\")\n",
        "X_cnn=pickle.load(pickle_in1)\n",
        "\n",
        "pickle_in2=open(\"Y_data_CNN.pickle\",\"rb\")\n",
        "Y_cnn=pickle.load(pickle_in2)"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HiakQU1YwShc"
      },
      "source": [
        "###Split into train and test sets\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lbUPQs8B9ocn",
        "outputId": "70acdf5a-101d-4bb8-cc39-c4b28ff24e54"
      },
      "source": [
        "X_trainn, X_testn, Y_train, Y_test = train_test_split(X_cnn , Y_cnn , test_size=0.15, shuffle=True, random_state=123, stratify=Y_cnn )\n",
        "\n",
        "#Normalize the image pixels to the range [0,1]\n",
        "X_train=X_trainn/255.0\n",
        "X_test=X_testn/255.0\n",
        "\n",
        "print(X_trainn.shape)\n",
        "print(Y_train.shape)\n",
        "print(X_testn.shape)\n",
        "print(Y_test.shape)"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(5124, 128, 128, 1)\n",
            "(5124,)\n",
            "(905, 128, 128, 1)\n",
            "(905,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XvStPgbclQjO"
      },
      "source": [
        "##Model configuration"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tMdRe6J8zfWF"
      },
      "source": [
        "cnn=models.Sequential([\n",
        "\n",
        "    layers.Conv2D(filters=64,kernel_size=(3,3),activation=\"relu\",input_shape=(X_train.shape[1:])),\n",
        "    layers.MaxPool2D(pool_size=(2,2)),\n",
        "    layers.Dropout(0.05),\n",
        "\n",
        "    layers.Conv2D(filters=64,kernel_size=(3,3),activation=\"relu\"),\n",
        "    layers.MaxPooling2D((2,2)),\n",
        "    layers.Dropout(0.07),\n",
        "\n",
        "    layers.Conv2D(filters=32,kernel_size=(3,3),activation=\"relu\"),\n",
        "    layers.MaxPooling2D((2,2)),\n",
        "    layers.Dropout(0.15),\n",
        "\n",
        "    #dense\n",
        "\n",
        "    layers.Flatten(name='flatten'),\n",
        "    layers.Dense(150,activation=\"relu\",name='dense1'),\n",
        "    layers.Dropout(0.15),\n",
        "    layers.Dense(7,activation=\"softmax\")\n",
        "    ])\n",
        "\n",
        "cnn.compile(optimizer='adam',\n",
        "            loss='sparse_categorical_crossentropy',\n",
        "            metrics=['accuracy'])"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Model Training"
      ],
      "metadata": {
        "id": "XGjZcciDrfXM"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "DSlXPmVE_TyX",
        "outputId": "28f62b28-484c-488c-84e0-5469b5eae40c"
      },
      "source": [
        "history=cnn.fit(X_train,Y_train,batch_size=30,validation_split=0.1,epochs=15)\n",
        "\n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['tain','val'],loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['tain','val'],loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/15\n",
            "154/154 [==============================] - 269s 2s/step - loss: 1.9248 - accuracy: 0.1739 - val_loss: 1.8984 - val_accuracy: 0.2378\n",
            "Epoch 2/15\n",
            "154/154 [==============================] - 253s 2s/step - loss: 1.3277 - accuracy: 0.4860 - val_loss: 0.9816 - val_accuracy: 0.6218\n",
            "Epoch 3/15\n",
            "154/154 [==============================] - 243s 2s/step - loss: 0.8590 - accuracy: 0.6825 - val_loss: 0.7667 - val_accuracy: 0.7310\n",
            "Epoch 4/15\n",
            "154/154 [==============================] - 246s 2s/step - loss: 0.6671 - accuracy: 0.7597 - val_loss: 0.6051 - val_accuracy: 0.7914\n",
            "Epoch 5/15\n",
            "154/154 [==============================] - 241s 2s/step - loss: 0.5564 - accuracy: 0.8057 - val_loss: 0.5527 - val_accuracy: 0.8207\n",
            "Epoch 6/15\n",
            "154/154 [==============================] - 250s 2s/step - loss: 0.4844 - accuracy: 0.8287 - val_loss: 0.5546 - val_accuracy: 0.8148\n",
            "Epoch 7/15\n",
            "154/154 [==============================] - 244s 2s/step - loss: 0.4193 - accuracy: 0.8508 - val_loss: 0.4723 - val_accuracy: 0.8441\n",
            "Epoch 8/15\n",
            "154/154 [==============================] - 241s 2s/step - loss: 0.3511 - accuracy: 0.8773 - val_loss: 0.4411 - val_accuracy: 0.8480\n",
            "Epoch 9/15\n",
            "154/154 [==============================] - 250s 2s/step - loss: 0.3209 - accuracy: 0.8842 - val_loss: 0.4672 - val_accuracy: 0.8558\n",
            "Epoch 10/15\n",
            "154/154 [==============================] - 249s 2s/step - loss: 0.3064 - accuracy: 0.8905 - val_loss: 0.4950 - val_accuracy: 0.8460\n",
            "Epoch 11/15\n",
            "154/154 [==============================] - 246s 2s/step - loss: 0.2662 - accuracy: 0.9074 - val_loss: 0.5191 - val_accuracy: 0.8499\n",
            "Epoch 12/15\n",
            "154/154 [==============================] - 242s 2s/step - loss: 0.2630 - accuracy: 0.9044 - val_loss: 0.4897 - val_accuracy: 0.8558\n",
            "Epoch 13/15\n",
            "154/154 [==============================] - 250s 2s/step - loss: 0.2490 - accuracy: 0.9111 - val_loss: 0.5113 - val_accuracy: 0.8519\n",
            "Epoch 14/15\n",
            "154/154 [==============================] - 243s 2s/step - loss: 0.2271 - accuracy: 0.9150 - val_loss: 0.4845 - val_accuracy: 0.8538\n",
            "Epoch 15/15\n",
            "154/154 [==============================] - 250s 2s/step - loss: 0.2250 - accuracy: 0.9230 - val_loss: 0.5157 - val_accuracy: 0.8596\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "drLCFhXngXoQ"
      },
      "source": [
        "cnn.save(\"CNNmodel.h5\")\n",
        "c=models.load_model('CNNmodel.h5')"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KT0kOzhnlcO8"
      },
      "source": [
        "##Model testing and evaluation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mGogyinaEE3a",
        "outputId": "b69b0421-8383-4064-f375-1731b2eb3c9e"
      },
      "source": [
        "cnn.evaluate(X_test,Y_test)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "29/29 [==============================] - 10s 332ms/step - loss: 0.5267 - accuracy: 0.8552\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.5266959071159363, 0.8552486300468445]"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VWbxl8dn5MAW",
        "outputId": "bd683171-d452-4fed-8392-bbc9b7c82664"
      },
      "source": [
        "y_prediction=cnn.predict(X_test)\n",
        "y_prediction = [np.argmax(element)for element in y_prediction ]\n",
        "\n",
        "print(classification_report(Y_test,y_prediction))\n",
        "print(confusion_matrix(Y_test , y_prediction))\n",
        "\n",
        "roc = roc_auc_score(Y_test, cnn.predict(X_test), multi_class ='ovr')\n",
        "kappa = cohen_kappa_score(Y_test, cnn.predict(X_test).argmax(axis=1))\n",
        "\n",
        "print(\"Roc score\", roc)\n",
        "print(\"Kappa score\", kappa)"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "29/29 [==============================] - 17s 590ms/step\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.88      0.87      0.88       148\n",
            "           1       0.84      0.88      0.86       138\n",
            "           2       0.83      0.87      0.85       143\n",
            "           3       0.87      0.83      0.85       135\n",
            "           4       0.70      0.71      0.70        89\n",
            "           5       0.87      0.85      0.86        86\n",
            "           6       0.94      0.90      0.92       166\n",
            "\n",
            "    accuracy                           0.86       905\n",
            "   macro avg       0.85      0.85      0.85       905\n",
            "weighted avg       0.86      0.86      0.86       905\n",
            "\n",
            "[[129   1   3   4   7   1   3]\n",
            " [  1 122   6   4   3   1   1]\n",
            " [  7   5 125   2   2   1   1]\n",
            " [  3   6  11 112   1   2   0]\n",
            " [  1   8   6   6  63   2   3]\n",
            " [  4   1   0   0   7  73   1]\n",
            " [  1   3   0   1   7   4 150]]\n",
            "29/29 [==============================] - 11s 397ms/step\n",
            "29/29 [==============================] - 9s 314ms/step\n",
            "Roc score 0.9811695281499756\n",
            "Kappa score 0.8298171347036388\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kTreQ_RzxnM_"
      },
      "source": [
        "#Long Short Term Memory (LSTM)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Data preparation"
      ],
      "metadata": {
        "id": "Beol6DrHxU2f"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SwRiaHjyGJCX"
      },
      "source": [
        "###Extract image sequences from the videos\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#For potential data augmentation\n",
        "\"\"\"def rotation(image, seed=777,rota=0):\n",
        "    random.seed(seed)\n",
        "    angle= rota\n",
        "    r_img = rotate(image, angle, mode='reflect', reshape=False, order=0)\n",
        "    return r_img\n",
        "\n",
        "def v_transl(image, seed=777,decalage=0):\n",
        "    random.seed(seed)\n",
        "    n_pixels = decalage\n",
        "    vtranslated_img = np.roll(image, n_pixels, axis=0)\n",
        "    return vtranslated_img\n",
        "\n",
        "def h_transl(image, seed=777,decalage=0):\n",
        "    random.seed(seed)\n",
        "    n_pixels = decalage\n",
        "    htranslated_img = np.roll(image, n_pixels, axis=1)\n",
        "    return htranslated_img\"\"\""
      ],
      "metadata": {
        "id": "m1p-2Jsf9Gam"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oaNUSfm353RO",
        "outputId": "2d8ac261-d60a-4993-c80e-23560a953da2"
      },
      "source": [
        "label2int = {\n",
        "    'dislike' : 0,\n",
        "    'like' : 1,\n",
        "    'ok' : 2,\n",
        "    'point' : 3,\n",
        "    'slide_1' : 4,\n",
        "    'slide_2' : 5,\n",
        "    'take_a_picture' : 6\n",
        "}\n",
        "Y_set = []\n",
        "X_set = []\n",
        "i=0\n",
        "\n",
        "for path_director in glob.glob('/content/drive/MyDrive/demo_room/*'):\n",
        "\n",
        "  label1 = path_director.split('/')[-1]\n",
        "  print(label1)\n",
        "  for image_path in glob.glob(path_director+'/*'):\n",
        "    tempx=[]\n",
        "    tempy=[]\n",
        "    cap=cv2.VideoCapture(image_path)\n",
        "    label = image_path.split('/')[-1].split('_')[0]\n",
        "    if label=='slide' or label=='take'  :\n",
        "      label=label+'_'+image_path.split('/')[-1].split('_')[1]\n",
        "      if label=='take_a':\n",
        "        label=label+'_'+image_path.split('/')[-1].split('_')[2]\n",
        "    nb=0\n",
        "    while cap.isOpened():\n",
        "\n",
        "      imageId=cap.get(1)\n",
        "      ret,image=cap.read()\n",
        "      if ret !=True:\n",
        "        break\n",
        "\n",
        "      if  imageId != 0 :\n",
        "        if imageId % 4 == 0:\n",
        "          nb=nb+1\n",
        "          image=cv2.cvtColor(image , cv2.COLOR_BGR2RGB)\n",
        "\n",
        "          if image.ndim==3:\n",
        "            #print(image_path)\n",
        "            image = cv2.GaussianBlur(image, (5,5) , 100)\n",
        "            red , green , blue = image[:,:,0] ,image[:,:,1] , image[:,:,2]\n",
        "\n",
        "\n",
        "            image_hsv = rgb2hsv(image)\n",
        "            hue = image_hsv[:,:,0]\n",
        "            sat = image_hsv[:,:,1]\n",
        "\n",
        "\n",
        "            image_ycbcr = rgb2ycbcr(image)\n",
        "            y = image_ycbcr[:,:,0]\n",
        "            cb = image_ycbcr[:,:,1]\n",
        "            cr = image_ycbcr[:,:,2]\n",
        "\n",
        "\n",
        "            mask = (red > 95) & (green > 40) & (blue > 20) & (red > green) & (red > blue) & (np.abs(red - green) > 15) & (cr > 135) & (cb > 85) & (y > 80) & (cr <= (1.5862*cb)+20) & (cr >= (0.3448*cb)+76.2069) & (cr >= (-4.5652*cb)+234.5652) & (cr <= (-1.15 * cb)+301.75) & (cr <= (-2.2857*cb)+432.85)\n",
        "            mask = mask.astype(np.uint8)\n",
        "            mask[mask == 1] = 255\n",
        "            np.sum(mask == 255)\n",
        "\n",
        "\n",
        "            kernel = np.ones((5,5), np.uint8)\n",
        "            mask1 = cv2.erode(mask , kernel , iterations=2)\n",
        "\n",
        "            kernel = np.ones((5,5), np.uint8)\n",
        "            image = cv2.dilate(mask1 , kernel , iterations=4)\n",
        "            #image=cv2.cvtColor(image,cv2.COLOR_GRAY2RGB)\n",
        "            image=resize(image,(128,128))\n",
        "\n",
        "            tempx.append(image)\n",
        "            tempy.append(label2int[label])\n",
        "\n",
        "    X_set.append(tempx)\n",
        "    Y_set.append(tempy)\n",
        "    i=i+1\n",
        "print(i)"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "subject_5\n",
            "subject_4\n",
            "subject_3\n",
            "subject_1\n",
            "subject_2\n",
            "705\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xkK-XvAXrPMy"
      },
      "source": [
        "###Data reshape and normalization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ygS975Lt2ors",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "42cccd30-22cd-4aff-8169-208e2d551537"
      },
      "source": [
        "X_set_array=[]\n",
        "Y_set_array=[]\n",
        "\n",
        "for fich in X_set:\n",
        "  X=np.array(fich).reshape(-1,128,128,1)\n",
        "  #normalization to the range [0,1]\n",
        "  X=X/255\n",
        "  X_set_array.append(X)\n",
        "\n",
        "for fich in Y_set:\n",
        "  Y=np.array(fich)\n",
        "  Y_set_array.append(Y)\n",
        "\n",
        "print(X_set_array[1].shape)\n",
        "print(Y_set_array[1].shape)"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(11, 128, 128, 1)\n",
            "(11,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9riJoCLZro4u"
      },
      "source": [
        "### Load the CNN dense1 layer weights"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VmdLfT74BLKf"
      },
      "source": [
        "c=models.load_model('CNNmodel.h5')\n",
        "\n",
        "#Get the cnn dense1 layer weights which will be used for feature extraction in the LSTM model\n",
        "modelfeatures= models.Model(\n",
        "    inputs=c.input,\n",
        "    outputs=c.get_layer('dense1').output\n",
        ")"
      ],
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bkj-EIf5r_gH"
      },
      "source": [
        "### Feature extraction for each image"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BFVnT_kD6T-9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fd5d774b-16ca-4aab-bb52-874c3019c606"
      },
      "source": [
        "trainfeaturesblock=[]\n",
        "for X in X_set_array:\n",
        "  trainf=modelfeatures.predict(X)\n",
        "  trainfeaturesblock.append(trainf)\n",
        "\n",
        "print(trainfeaturesblock[0].shape,trainfeaturesblock[1].shape,trainfeaturesblock[2].shape)\n",
        "print(len(trainfeaturesblock))"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 397ms/step\n",
            "1/1 [==============================] - 0s 418ms/step\n",
            "1/1 [==============================] - 1s 679ms/step\n",
            "1/1 [==============================] - 0s 468ms/step\n",
            "1/1 [==============================] - 0s 471ms/step\n",
            "1/1 [==============================] - 0s 287ms/step\n",
            "1/1 [==============================] - 0s 320ms/step\n",
            "1/1 [==============================] - 0s 380ms/step\n",
            "1/1 [==============================] - 0s 311ms/step\n",
            "1/1 [==============================] - 0s 318ms/step\n",
            "1/1 [==============================] - 0s 239ms/step\n",
            "1/1 [==============================] - 0s 257ms/step\n",
            "1/1 [==============================] - 0s 304ms/step\n",
            "1/1 [==============================] - 0s 295ms/step\n",
            "1/1 [==============================] - 0s 247ms/step\n",
            "1/1 [==============================] - 0s 223ms/step\n",
            "1/1 [==============================] - 0s 286ms/step\n",
            "1/1 [==============================] - 0s 216ms/step\n",
            "1/1 [==============================] - 0s 267ms/step\n",
            "1/1 [==============================] - 0s 375ms/step\n",
            "1/1 [==============================] - 0s 170ms/step\n",
            "1/1 [==============================] - 0s 253ms/step\n",
            "1/1 [==============================] - 0s 247ms/step\n",
            "1/1 [==============================] - 0s 201ms/step\n",
            "1/1 [==============================] - 0s 245ms/step\n",
            "1/1 [==============================] - 0s 227ms/step\n",
            "1/1 [==============================] - 0s 236ms/step\n",
            "1/1 [==============================] - 0s 250ms/step\n",
            "1/1 [==============================] - 0s 263ms/step\n",
            "1/1 [==============================] - 0s 243ms/step\n",
            "1/1 [==============================] - 0s 238ms/step\n",
            "1/1 [==============================] - 0s 227ms/step\n",
            "1/1 [==============================] - 0s 214ms/step\n",
            "1/1 [==============================] - 0s 228ms/step\n",
            "1/1 [==============================] - 0s 268ms/step\n",
            "1/1 [==============================] - 0s 155ms/step\n",
            "1/1 [==============================] - 0s 220ms/step\n",
            "1/1 [==============================] - 0s 246ms/step\n",
            "1/1 [==============================] - 0s 323ms/step\n",
            "1/1 [==============================] - 0s 411ms/step\n",
            "1/1 [==============================] - 0s 333ms/step\n",
            "1/1 [==============================] - 0s 328ms/step\n",
            "1/1 [==============================] - 0s 327ms/step\n",
            "1/1 [==============================] - 0s 398ms/step\n",
            "1/1 [==============================] - 0s 415ms/step\n",
            "1/1 [==============================] - 0s 260ms/step\n",
            "1/1 [==============================] - 0s 418ms/step\n",
            "1/1 [==============================] - 0s 488ms/step\n",
            "1/1 [==============================] - 0s 451ms/step\n",
            "1/1 [==============================] - 0s 295ms/step\n",
            "1/1 [==============================] - 0s 228ms/step\n",
            "1/1 [==============================] - 0s 239ms/step\n",
            "1/1 [==============================] - 0s 225ms/step\n",
            "1/1 [==============================] - 0s 239ms/step\n",
            "1/1 [==============================] - 0s 255ms/step\n",
            "1/1 [==============================] - 0s 228ms/step\n",
            "1/1 [==============================] - 0s 175ms/step\n",
            "1/1 [==============================] - 0s 248ms/step\n",
            "1/1 [==============================] - 0s 257ms/step\n",
            "1/1 [==============================] - 0s 176ms/step\n",
            "1/1 [==============================] - 0s 170ms/step\n",
            "1/1 [==============================] - 0s 213ms/step\n",
            "1/1 [==============================] - 0s 156ms/step\n",
            "1/1 [==============================] - 0s 202ms/step\n",
            "1/1 [==============================] - 0s 195ms/step\n",
            "1/1 [==============================] - 0s 240ms/step\n",
            "1/1 [==============================] - 0s 209ms/step\n",
            "1/1 [==============================] - 0s 241ms/step\n",
            "1/1 [==============================] - 0s 169ms/step\n",
            "1/1 [==============================] - 0s 229ms/step\n",
            "1/1 [==============================] - 0s 183ms/step\n",
            "1/1 [==============================] - 0s 167ms/step\n",
            "1/1 [==============================] - 0s 161ms/step\n",
            "1/1 [==============================] - 0s 220ms/step\n",
            "1/1 [==============================] - 0s 233ms/step\n",
            "1/1 [==============================] - 0s 270ms/step\n",
            "1/1 [==============================] - 0s 164ms/step\n",
            "1/1 [==============================] - 0s 231ms/step\n",
            "1/1 [==============================] - 0s 271ms/step\n",
            "1/1 [==============================] - 0s 222ms/step\n",
            "1/1 [==============================] - 0s 228ms/step\n",
            "1/1 [==============================] - 0s 187ms/step\n",
            "1/1 [==============================] - 0s 330ms/step\n",
            "1/1 [==============================] - 0s 403ms/step\n",
            "1/1 [==============================] - 0s 374ms/step\n",
            "1/1 [==============================] - 0s 402ms/step\n",
            "1/1 [==============================] - 0s 322ms/step\n",
            "1/1 [==============================] - 0s 333ms/step\n",
            "1/1 [==============================] - 0s 419ms/step\n",
            "1/1 [==============================] - 0s 200ms/step\n",
            "1/1 [==============================] - 0s 346ms/step\n",
            "1/1 [==============================] - 0s 431ms/step\n",
            "1/1 [==============================] - 0s 407ms/step\n",
            "1/1 [==============================] - 0s 382ms/step\n",
            "1/1 [==============================] - 0s 174ms/step\n",
            "1/1 [==============================] - 0s 207ms/step\n",
            "1/1 [==============================] - 0s 169ms/step\n",
            "1/1 [==============================] - 0s 210ms/step\n",
            "1/1 [==============================] - 0s 246ms/step\n",
            "1/1 [==============================] - 0s 227ms/step\n",
            "1/1 [==============================] - 0s 223ms/step\n",
            "1/1 [==============================] - 0s 216ms/step\n",
            "1/1 [==============================] - 0s 200ms/step\n",
            "1/1 [==============================] - 0s 190ms/step\n",
            "1/1 [==============================] - 0s 273ms/step\n",
            "1/1 [==============================] - 0s 170ms/step\n",
            "1/1 [==============================] - 0s 141ms/step\n",
            "1/1 [==============================] - 0s 152ms/step\n",
            "1/1 [==============================] - 0s 154ms/step\n",
            "1/1 [==============================] - 0s 144ms/step\n",
            "1/1 [==============================] - 0s 199ms/step\n",
            "1/1 [==============================] - 0s 153ms/step\n",
            "1/1 [==============================] - 0s 147ms/step\n",
            "1/1 [==============================] - 0s 184ms/step\n",
            "1/1 [==============================] - 0s 157ms/step\n",
            "1/1 [==============================] - 0s 143ms/step\n",
            "1/1 [==============================] - 0s 137ms/step\n",
            "1/1 [==============================] - 0s 156ms/step\n",
            "1/1 [==============================] - 0s 147ms/step\n",
            "1/1 [==============================] - 0s 166ms/step\n",
            "1/1 [==============================] - 0s 143ms/step\n",
            "1/1 [==============================] - 0s 133ms/step\n",
            "1/1 [==============================] - 0s 135ms/step\n",
            "1/1 [==============================] - 0s 137ms/step\n",
            "1/1 [==============================] - 0s 181ms/step\n",
            "1/1 [==============================] - 0s 179ms/step\n",
            "1/1 [==============================] - 0s 119ms/step\n",
            "1/1 [==============================] - 0s 131ms/step\n",
            "1/1 [==============================] - 0s 159ms/step\n",
            "1/1 [==============================] - 0s 146ms/step\n",
            "1/1 [==============================] - 0s 123ms/step\n",
            "1/1 [==============================] - 0s 145ms/step\n",
            "1/1 [==============================] - 0s 118ms/step\n",
            "1/1 [==============================] - 0s 180ms/step\n",
            "1/1 [==============================] - 0s 175ms/step\n",
            "1/1 [==============================] - 0s 289ms/step\n",
            "1/1 [==============================] - 0s 276ms/step\n",
            "1/1 [==============================] - 0s 232ms/step\n",
            "1/1 [==============================] - 0s 203ms/step\n",
            "1/1 [==============================] - 0s 245ms/step\n",
            "1/1 [==============================] - 0s 192ms/step\n",
            "1/1 [==============================] - 0s 209ms/step\n",
            "1/1 [==============================] - 0s 155ms/step\n",
            "1/1 [==============================] - 0s 222ms/step\n",
            "1/1 [==============================] - 0s 203ms/step\n",
            "1/1 [==============================] - 0s 323ms/step\n",
            "1/1 [==============================] - 0s 382ms/step\n",
            "1/1 [==============================] - 0s 410ms/step\n",
            "1/1 [==============================] - 0s 330ms/step\n",
            "1/1 [==============================] - 0s 292ms/step\n",
            "1/1 [==============================] - 0s 235ms/step\n",
            "1/1 [==============================] - 0s 251ms/step\n",
            "1/1 [==============================] - 0s 211ms/step\n",
            "1/1 [==============================] - 0s 185ms/step\n",
            "1/1 [==============================] - 0s 280ms/step\n",
            "1/1 [==============================] - 0s 160ms/step\n",
            "1/1 [==============================] - 0s 303ms/step\n",
            "1/1 [==============================] - 0s 264ms/step\n",
            "1/1 [==============================] - 0s 254ms/step\n",
            "1/1 [==============================] - 0s 196ms/step\n",
            "1/1 [==============================] - 0s 228ms/step\n",
            "1/1 [==============================] - 0s 259ms/step\n",
            "1/1 [==============================] - 0s 212ms/step\n",
            "1/1 [==============================] - 0s 208ms/step\n",
            "1/1 [==============================] - 0s 263ms/step\n",
            "1/1 [==============================] - 0s 195ms/step\n",
            "1/1 [==============================] - 0s 216ms/step\n",
            "1/1 [==============================] - 0s 132ms/step\n",
            "1/1 [==============================] - 0s 124ms/step\n",
            "1/1 [==============================] - 0s 140ms/step\n",
            "1/1 [==============================] - 0s 105ms/step\n",
            "1/1 [==============================] - 0s 139ms/step\n",
            "1/1 [==============================] - 0s 145ms/step\n",
            "1/1 [==============================] - 0s 140ms/step\n",
            "1/1 [==============================] - 0s 164ms/step\n",
            "1/1 [==============================] - 0s 99ms/step\n",
            "1/1 [==============================] - 0s 133ms/step\n",
            "1/1 [==============================] - 0s 150ms/step\n",
            "1/1 [==============================] - 0s 128ms/step\n",
            "1/1 [==============================] - 0s 157ms/step\n",
            "1/1 [==============================] - 0s 134ms/step\n",
            "1/1 [==============================] - 0s 132ms/step\n",
            "1/1 [==============================] - 0s 96ms/step\n",
            "1/1 [==============================] - 0s 144ms/step\n",
            "1/1 [==============================] - 0s 128ms/step\n",
            "1/1 [==============================] - 0s 112ms/step\n",
            "1/1 [==============================] - 0s 104ms/step\n",
            "1/1 [==============================] - 0s 136ms/step\n",
            "1/1 [==============================] - 0s 156ms/step\n",
            "1/1 [==============================] - 0s 128ms/step\n",
            "1/1 [==============================] - 0s 101ms/step\n",
            "1/1 [==============================] - 0s 206ms/step\n",
            "1/1 [==============================] - 0s 203ms/step\n",
            "1/1 [==============================] - 0s 218ms/step\n",
            "1/1 [==============================] - 0s 279ms/step\n",
            "1/1 [==============================] - 0s 217ms/step\n",
            "1/1 [==============================] - 0s 205ms/step\n",
            "1/1 [==============================] - 0s 191ms/step\n",
            "1/1 [==============================] - 0s 253ms/step\n",
            "1/1 [==============================] - 0s 218ms/step\n",
            "1/1 [==============================] - 0s 213ms/step\n",
            "1/1 [==============================] - 0s 231ms/step\n",
            "1/1 [==============================] - 0s 244ms/step\n",
            "1/1 [==============================] - 0s 228ms/step\n",
            "1/1 [==============================] - 0s 306ms/step\n",
            "1/1 [==============================] - 0s 232ms/step\n",
            "1/1 [==============================] - 0s 227ms/step\n",
            "1/1 [==============================] - 0s 158ms/step\n",
            "1/1 [==============================] - 0s 169ms/step\n",
            "1/1 [==============================] - 0s 152ms/step\n",
            "1/1 [==============================] - 0s 129ms/step\n",
            "1/1 [==============================] - 0s 182ms/step\n",
            "1/1 [==============================] - 0s 132ms/step\n",
            "1/1 [==============================] - 0s 117ms/step\n",
            "1/1 [==============================] - 0s 162ms/step\n",
            "1/1 [==============================] - 0s 143ms/step\n",
            "1/1 [==============================] - 0s 132ms/step\n",
            "1/1 [==============================] - 0s 127ms/step\n",
            "1/1 [==============================] - 0s 144ms/step\n",
            "1/1 [==============================] - 0s 148ms/step\n",
            "1/1 [==============================] - 0s 139ms/step\n",
            "1/1 [==============================] - 0s 129ms/step\n",
            "1/1 [==============================] - 0s 103ms/step\n",
            "1/1 [==============================] - 0s 145ms/step\n",
            "1/1 [==============================] - 0s 198ms/step\n",
            "1/1 [==============================] - 0s 138ms/step\n",
            "1/1 [==============================] - 0s 138ms/step\n",
            "1/1 [==============================] - 0s 127ms/step\n",
            "1/1 [==============================] - 0s 127ms/step\n",
            "1/1 [==============================] - 0s 197ms/step\n",
            "1/1 [==============================] - 0s 103ms/step\n",
            "1/1 [==============================] - 0s 141ms/step\n",
            "1/1 [==============================] - 0s 135ms/step\n",
            "1/1 [==============================] - 0s 148ms/step\n",
            "1/1 [==============================] - 0s 327ms/step\n",
            "1/1 [==============================] - 0s 318ms/step\n",
            "1/1 [==============================] - 0s 267ms/step\n",
            "1/1 [==============================] - 0s 265ms/step\n",
            "1/1 [==============================] - 0s 137ms/step\n",
            "1/1 [==============================] - 0s 150ms/step\n",
            "1/1 [==============================] - 0s 155ms/step\n",
            "1/1 [==============================] - 0s 131ms/step\n",
            "1/1 [==============================] - 0s 102ms/step\n",
            "1/1 [==============================] - 0s 123ms/step\n",
            "1/1 [==============================] - 0s 141ms/step\n",
            "1/1 [==============================] - 0s 61ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 56ms/step\n",
            "1/1 [==============================] - 0s 55ms/step\n",
            "1/1 [==============================] - 0s 56ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "1/1 [==============================] - 0s 60ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 112ms/step\n",
            "1/1 [==============================] - 0s 113ms/step\n",
            "1/1 [==============================] - 0s 107ms/step\n",
            "1/1 [==============================] - 0s 122ms/step\n",
            "1/1 [==============================] - 0s 115ms/step\n",
            "1/1 [==============================] - 0s 99ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 101ms/step\n",
            "1/1 [==============================] - 0s 64ms/step\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "1/1 [==============================] - 0s 110ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "1/1 [==============================] - 0s 104ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 100ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 179ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 142ms/step\n",
            "1/1 [==============================] - 0s 131ms/step\n",
            "1/1 [==============================] - 0s 149ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 95ms/step\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 96ms/step\n",
            "1/1 [==============================] - 0s 88ms/step\n",
            "1/1 [==============================] - 0s 110ms/step\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "1/1 [==============================] - 0s 95ms/step\n",
            "1/1 [==============================] - 0s 111ms/step\n",
            "1/1 [==============================] - 0s 114ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 88ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 104ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 122ms/step\n",
            "1/1 [==============================] - 0s 121ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 83ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 120ms/step\n",
            "1/1 [==============================] - 0s 93ms/step\n",
            "1/1 [==============================] - 0s 100ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 100ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 122ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 90ms/step\n",
            "1/1 [==============================] - 0s 119ms/step\n",
            "1/1 [==============================] - 0s 124ms/step\n",
            "1/1 [==============================] - 0s 90ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 88ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 96ms/step\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 94ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 145ms/step\n",
            "1/1 [==============================] - 0s 128ms/step\n",
            "1/1 [==============================] - 0s 150ms/step\n",
            "1/1 [==============================] - 0s 159ms/step\n",
            "1/1 [==============================] - 0s 123ms/step\n",
            "1/1 [==============================] - 0s 138ms/step\n",
            "1/1 [==============================] - 0s 124ms/step\n",
            "1/1 [==============================] - 0s 163ms/step\n",
            "1/1 [==============================] - 0s 152ms/step\n",
            "1/1 [==============================] - 0s 107ms/step\n",
            "1/1 [==============================] - 0s 142ms/step\n",
            "1/1 [==============================] - 0s 138ms/step\n",
            "1/1 [==============================] - 0s 119ms/step\n",
            "1/1 [==============================] - 0s 113ms/step\n",
            "1/1 [==============================] - 0s 139ms/step\n",
            "1/1 [==============================] - 0s 118ms/step\n",
            "1/1 [==============================] - 0s 141ms/step\n",
            "1/1 [==============================] - 0s 134ms/step\n",
            "1/1 [==============================] - 0s 147ms/step\n",
            "1/1 [==============================] - 0s 112ms/step\n",
            "1/1 [==============================] - 0s 97ms/step\n",
            "1/1 [==============================] - 0s 61ms/step\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 61ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 98ms/step\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "1/1 [==============================] - 0s 118ms/step\n",
            "1/1 [==============================] - 0s 110ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 61ms/step\n",
            "1/1 [==============================] - 0s 55ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 65ms/step\n",
            "1/1 [==============================] - 0s 98ms/step\n",
            "1/1 [==============================] - 0s 83ms/step\n",
            "1/1 [==============================] - 0s 93ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "1/1 [==============================] - 0s 88ms/step\n",
            "1/1 [==============================] - 0s 126ms/step\n",
            "1/1 [==============================] - 0s 175ms/step\n",
            "1/1 [==============================] - 0s 134ms/step\n",
            "1/1 [==============================] - 0s 96ms/step\n",
            "1/1 [==============================] - 0s 95ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 103ms/step\n",
            "1/1 [==============================] - 0s 103ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 88ms/step\n",
            "1/1 [==============================] - 0s 99ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 126ms/step\n",
            "1/1 [==============================] - 0s 123ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 220ms/step\n",
            "1/1 [==============================] - 0s 159ms/step\n",
            "1/1 [==============================] - 0s 187ms/step\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "1/1 [==============================] - 0s 105ms/step\n",
            "1/1 [==============================] - 0s 118ms/step\n",
            "1/1 [==============================] - 0s 121ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 131ms/step\n",
            "1/1 [==============================] - 0s 102ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 143ms/step\n",
            "1/1 [==============================] - 0s 109ms/step\n",
            "1/1 [==============================] - 0s 100ms/step\n",
            "1/1 [==============================] - 0s 100ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 248ms/step\n",
            "1/1 [==============================] - 0s 152ms/step\n",
            "1/1 [==============================] - 0s 210ms/step\n",
            "1/1 [==============================] - 0s 193ms/step\n",
            "1/1 [==============================] - 0s 139ms/step\n",
            "1/1 [==============================] - 0s 122ms/step\n",
            "1/1 [==============================] - 0s 274ms/step\n",
            "1/1 [==============================] - 0s 222ms/step\n",
            "1/1 [==============================] - 0s 154ms/step\n",
            "1/1 [==============================] - 0s 210ms/step\n",
            "1/1 [==============================] - 0s 128ms/step\n",
            "1/1 [==============================] - 0s 172ms/step\n",
            "1/1 [==============================] - 0s 196ms/step\n",
            "1/1 [==============================] - 0s 154ms/step\n",
            "1/1 [==============================] - 0s 226ms/step\n",
            "1/1 [==============================] - 0s 234ms/step\n",
            "1/1 [==============================] - 0s 143ms/step\n",
            "1/1 [==============================] - 0s 185ms/step\n",
            "1/1 [==============================] - 0s 182ms/step\n",
            "1/1 [==============================] - 0s 311ms/step\n",
            "1/1 [==============================] - 0s 172ms/step\n",
            "1/1 [==============================] - 0s 174ms/step\n",
            "1/1 [==============================] - 0s 212ms/step\n",
            "1/1 [==============================] - 0s 136ms/step\n",
            "1/1 [==============================] - 0s 267ms/step\n",
            "1/1 [==============================] - 0s 244ms/step\n",
            "1/1 [==============================] - 0s 169ms/step\n",
            "1/1 [==============================] - 0s 212ms/step\n",
            "1/1 [==============================] - 0s 379ms/step\n",
            "1/1 [==============================] - 0s 223ms/step\n",
            "1/1 [==============================] - 0s 274ms/step\n",
            "1/1 [==============================] - 0s 234ms/step\n",
            "1/1 [==============================] - 0s 204ms/step\n",
            "1/1 [==============================] - 0s 103ms/step\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 156ms/step\n",
            "1/1 [==============================] - 0s 155ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 103ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 177ms/step\n",
            "1/1 [==============================] - 0s 106ms/step\n",
            "1/1 [==============================] - 0s 159ms/step\n",
            "1/1 [==============================] - 0s 127ms/step\n",
            "1/1 [==============================] - 0s 158ms/step\n",
            "1/1 [==============================] - 0s 231ms/step\n",
            "1/1 [==============================] - 0s 106ms/step\n",
            "1/1 [==============================] - 0s 97ms/step\n",
            "1/1 [==============================] - 0s 141ms/step\n",
            "1/1 [==============================] - 0s 130ms/step\n",
            "1/1 [==============================] - 0s 154ms/step\n",
            "1/1 [==============================] - 0s 147ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 133ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 110ms/step\n",
            "1/1 [==============================] - 0s 120ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 170ms/step\n",
            "1/1 [==============================] - 0s 125ms/step\n",
            "1/1 [==============================] - 0s 139ms/step\n",
            "1/1 [==============================] - 0s 152ms/step\n",
            "1/1 [==============================] - 0s 225ms/step\n",
            "1/1 [==============================] - 0s 147ms/step\n",
            "1/1 [==============================] - 0s 130ms/step\n",
            "1/1 [==============================] - 0s 332ms/step\n",
            "1/1 [==============================] - 0s 147ms/step\n",
            "1/1 [==============================] - 0s 113ms/step\n",
            "1/1 [==============================] - 0s 124ms/step\n",
            "1/1 [==============================] - 0s 167ms/step\n",
            "1/1 [==============================] - 0s 137ms/step\n",
            "1/1 [==============================] - 0s 135ms/step\n",
            "1/1 [==============================] - 0s 271ms/step\n",
            "1/1 [==============================] - 0s 211ms/step\n",
            "1/1 [==============================] - 0s 164ms/step\n",
            "1/1 [==============================] - 0s 119ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 118ms/step\n",
            "1/1 [==============================] - 0s 108ms/step\n",
            "1/1 [==============================] - 0s 143ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 110ms/step\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "1/1 [==============================] - 0s 120ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 105ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 109ms/step\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 141ms/step\n",
            "1/1 [==============================] - 0s 152ms/step\n",
            "1/1 [==============================] - 0s 97ms/step\n",
            "1/1 [==============================] - 0s 171ms/step\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 90ms/step\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "1/1 [==============================] - 0s 188ms/step\n",
            "1/1 [==============================] - 0s 135ms/step\n",
            "1/1 [==============================] - 0s 130ms/step\n",
            "1/1 [==============================] - 0s 168ms/step\n",
            "1/1 [==============================] - 0s 187ms/step\n",
            "1/1 [==============================] - 0s 132ms/step\n",
            "1/1 [==============================] - 0s 131ms/step\n",
            "1/1 [==============================] - 0s 100ms/step\n",
            "1/1 [==============================] - 0s 130ms/step\n",
            "1/1 [==============================] - 0s 148ms/step\n",
            "1/1 [==============================] - 0s 127ms/step\n",
            "1/1 [==============================] - 0s 121ms/step\n",
            "1/1 [==============================] - 0s 161ms/step\n",
            "1/1 [==============================] - 0s 130ms/step\n",
            "1/1 [==============================] - 0s 145ms/step\n",
            "1/1 [==============================] - 0s 121ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 60ms/step\n",
            "1/1 [==============================] - 0s 141ms/step\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "1/1 [==============================] - 0s 105ms/step\n",
            "1/1 [==============================] - 0s 99ms/step\n",
            "1/1 [==============================] - 0s 98ms/step\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "1/1 [==============================] - 0s 90ms/step\n",
            "1/1 [==============================] - 0s 75ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 102ms/step\n",
            "1/1 [==============================] - 0s 56ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 83ms/step\n",
            "1/1 [==============================] - 0s 75ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 56ms/step\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 99ms/step\n",
            "1/1 [==============================] - 0s 210ms/step\n",
            "1/1 [==============================] - 0s 139ms/step\n",
            "1/1 [==============================] - 0s 160ms/step\n",
            "1/1 [==============================] - 0s 83ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 138ms/step\n",
            "1/1 [==============================] - 0s 120ms/step\n",
            "1/1 [==============================] - 0s 151ms/step\n",
            "1/1 [==============================] - 0s 162ms/step\n",
            "1/1 [==============================] - 0s 134ms/step\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 160ms/step\n",
            "1/1 [==============================] - 0s 171ms/step\n",
            "1/1 [==============================] - 0s 147ms/step\n",
            "1/1 [==============================] - 0s 168ms/step\n",
            "1/1 [==============================] - 0s 155ms/step\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 145ms/step\n",
            "1/1 [==============================] - 0s 114ms/step\n",
            "1/1 [==============================] - 0s 305ms/step\n",
            "1/1 [==============================] - 0s 163ms/step\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "1/1 [==============================] - 0s 98ms/step\n",
            "1/1 [==============================] - 0s 56ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 94ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 65ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 106ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 61ms/step\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "1/1 [==============================] - 0s 127ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 65ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 94ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 116ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 86ms/step\n",
            "1/1 [==============================] - 0s 114ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 114ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 64ms/step\n",
            "1/1 [==============================] - 0s 64ms/step\n",
            "1/1 [==============================] - 0s 95ms/step\n",
            "1/1 [==============================] - 0s 65ms/step\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 85ms/step\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 110ms/step\n",
            "1/1 [==============================] - 0s 172ms/step\n",
            "1/1 [==============================] - 0s 89ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 125ms/step\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "1/1 [==============================] - 0s 123ms/step\n",
            "1/1 [==============================] - 0s 125ms/step\n",
            "1/1 [==============================] - 0s 157ms/step\n",
            "1/1 [==============================] - 0s 119ms/step\n",
            "1/1 [==============================] - 0s 151ms/step\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "1/1 [==============================] - 0s 64ms/step\n",
            "1/1 [==============================] - 0s 131ms/step\n",
            "1/1 [==============================] - 0s 112ms/step\n",
            "1/1 [==============================] - 0s 162ms/step\n",
            "1/1 [==============================] - 0s 213ms/step\n",
            "(10, 150) (11, 150) (12, 150)\n",
            "705\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1mlFKmNlsJMb"
      },
      "source": [
        "### Put the images in sequences of 6"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_XIey-M2TBvI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "79e7b104-156d-4d52-cd5d-a0bfb22cd2a5"
      },
      "source": [
        "X_data=[]\n",
        "for X in trainfeaturesblock:\n",
        "  li=[]\n",
        "  i=0\n",
        "  for ph in X:\n",
        "    i=i+1\n",
        "    li.append(ph)\n",
        "    if i==6:\n",
        "      X_data.append(li)\n",
        "      i=0\n",
        "      li=[]\n",
        "\n",
        "print(len(X_data))"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "721\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6"
            ]
          },
          "metadata": {},
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Reshape data\n"
      ],
      "metadata": {
        "id": "vw2tB_quwvdR"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1cG_1-zFkGPs",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1159b045-0787-43f6-bae8-c449ed535b39"
      },
      "source": [
        "Y_data1=[]\n",
        "for Y in Y_set_array:\n",
        "  li=[]\n",
        "  i=0\n",
        "  for classe in Y:\n",
        "    i=i+1\n",
        "    li.append(classe)\n",
        "    if i==6:\n",
        "      Y_data1.append(li)\n",
        "      i=0\n",
        "      li=[]\n",
        "\n",
        "print(len(Y_data1))\n",
        "print(len(Y_data1[0]))"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "721\n",
            "6\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mxhz7s_YgQVq",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "15f6c6c0-d9f4-47cb-ac66-acca9bda2765"
      },
      "source": [
        "Y_data=[]\n",
        "for Y in Y_data1:\n",
        "  Y_data.append(Y[0])\n",
        "len(Y_data)"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "721"
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vKmE6X72lk46",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cb8c9ab3-cd71-4ffa-b36c-415a01b0cd83"
      },
      "source": [
        "X_dataset=np.array(X_data).reshape(-1,6,150)\n",
        "Y_dataset=np.array(Y_data)\n",
        "\n",
        "print(X_dataset.shape)\n",
        "print(Y_dataset.shape)"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(721, 6, 150)\n",
            "(721,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Save the obtained data"
      ],
      "metadata": {
        "id": "n-fFSf_Spgiq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "pickle_out=open(\"X_data_LSTM.pickle\",\"wb\")\n",
        "pickle.dump(X_dataset,pickle_out)\n",
        "pickle_out.close()\n",
        "\n",
        "pickle_out=open(\"Y_data_LSTM.pickle\",\"wb\")\n",
        "pickle.dump(Y_dataset,pickle_out)\n",
        "pickle_out.close()"
      ],
      "metadata": {
        "id": "_6FHhqrKo9hj"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "pickle_in1=open(\"X_data_LSTM.pickle\",\"rb\")\n",
        "X_dataset=pickle.load(pickle_in1)\n",
        "\n",
        "pickle_in2=open(\"Y_data_LSTM.pickle\",\"rb\")\n",
        "Y_dataset=pickle.load(pickle_in2)"
      ],
      "metadata": {
        "id": "VI7Ld9onprRt"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Split into train and test sets\n"
      ],
      "metadata": {
        "id": "1dTQ9vUyx8L4"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qLozfhEcoChf",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9eebda50-9647-499b-93e7-3d99187b267b"
      },
      "source": [
        "X_train, X_test, Y_train, Y_test = train_test_split(X_dataset , Y_dataset , test_size=0.15, random_state=123,stratify=Y_dataset)\n",
        "\n",
        "print(X_train.shape)\n",
        "print(Y_train.shape)\n",
        "print(X_test.shape)\n",
        "print(Y_test.shape)"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(612, 6, 150)\n",
            "(612,)\n",
            "(109, 6, 150)\n",
            "(109,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_u49pHnm8Vvg"
      },
      "source": [
        "##Model configuration"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DzHJnX1Cxt2Y"
      },
      "source": [
        "lstm=models.Sequential([\n",
        "\n",
        "layers.LSTM(128,input_shape=(X_train.shape[1:]),activation='relu',return_sequences=True),\n",
        "layers.Dropout(0.2),\n",
        "\n",
        "layers.LSTM(128,activation='relu'),\n",
        "layers.Dropout(0.2),\n",
        "\n",
        "layers.Dense(32,activation='relu'),\n",
        "layers.Dropout(0.2),\n",
        "\n",
        "layers.Dense(7,activation='softmax'),\n",
        "\n",
        "])\n",
        "\n",
        "opt= tensorflow.keras.optimizers.Adam(learning_rate=1e-4)\n",
        "\n",
        "lstm.compile(loss='sparse_categorical_crossentropy',optimizer=opt,metrics=['accuracy'])"
      ],
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Model training"
      ],
      "metadata": {
        "id": "6-nYXrwHycB1"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "yHyAh7UM5249",
        "outputId": "50907291-ce93-4561-d8ef-f0dd48af2758"
      },
      "source": [
        "history1=lstm.fit(X_train,Y_train,batch_size=30,validation_split=0.15,epochs=30)\n",
        "\n",
        "plt.plot(history1.history['accuracy'])\n",
        "plt.plot(history1.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['tain','val'],loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(history1.history['loss'])\n",
        "plt.plot(history1.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['tain','val'],loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/30\n",
            "18/18 [==============================] - 5s 70ms/step - loss: 1.9919 - accuracy: 0.1942 - val_loss: 1.8261 - val_accuracy: 0.2826\n",
            "Epoch 2/30\n",
            "18/18 [==============================] - 1s 54ms/step - loss: 1.7181 - accuracy: 0.3462 - val_loss: 1.6890 - val_accuracy: 0.4239\n",
            "Epoch 3/30\n",
            "18/18 [==============================] - 1s 54ms/step - loss: 1.5622 - accuracy: 0.4442 - val_loss: 1.5517 - val_accuracy: 0.5109\n",
            "Epoch 4/30\n",
            "18/18 [==============================] - 1s 59ms/step - loss: 1.4024 - accuracy: 0.5115 - val_loss: 1.3902 - val_accuracy: 0.5000\n",
            "Epoch 5/30\n",
            "18/18 [==============================] - 1s 59ms/step - loss: 1.2887 - accuracy: 0.5269 - val_loss: 1.1978 - val_accuracy: 0.5761\n",
            "Epoch 6/30\n",
            "18/18 [==============================] - 1s 44ms/step - loss: 1.0900 - accuracy: 0.5788 - val_loss: 0.9998 - val_accuracy: 0.5870\n",
            "Epoch 7/30\n",
            "18/18 [==============================] - 1s 31ms/step - loss: 0.9582 - accuracy: 0.6019 - val_loss: 0.8159 - val_accuracy: 0.6630\n",
            "Epoch 8/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.7789 - accuracy: 0.6731 - val_loss: 0.6346 - val_accuracy: 0.8370\n",
            "Epoch 9/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.6724 - accuracy: 0.7327 - val_loss: 0.4507 - val_accuracy: 0.9348\n",
            "Epoch 10/30\n",
            "18/18 [==============================] - 1s 31ms/step - loss: 0.4722 - accuracy: 0.8288 - val_loss: 0.2610 - val_accuracy: 0.9565\n",
            "Epoch 11/30\n",
            "18/18 [==============================] - 1s 36ms/step - loss: 0.3343 - accuracy: 0.8885 - val_loss: 0.1236 - val_accuracy: 1.0000\n",
            "Epoch 12/30\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 0.2489 - accuracy: 0.9231 - val_loss: 0.0848 - val_accuracy: 0.9891\n",
            "Epoch 13/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.2178 - accuracy: 0.9269 - val_loss: 0.0602 - val_accuracy: 1.0000\n",
            "Epoch 14/30\n",
            "18/18 [==============================] - 1s 32ms/step - loss: 0.1534 - accuracy: 0.9500 - val_loss: 0.0465 - val_accuracy: 1.0000\n",
            "Epoch 15/30\n",
            "18/18 [==============================] - 1s 32ms/step - loss: 0.1306 - accuracy: 0.9500 - val_loss: 0.0360 - val_accuracy: 1.0000\n",
            "Epoch 16/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.1291 - accuracy: 0.9615 - val_loss: 0.0270 - val_accuracy: 1.0000\n",
            "Epoch 17/30\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 0.1012 - accuracy: 0.9615 - val_loss: 0.0246 - val_accuracy: 1.0000\n",
            "Epoch 18/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.0807 - accuracy: 0.9750 - val_loss: 0.0215 - val_accuracy: 1.0000\n",
            "Epoch 19/30\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 0.0735 - accuracy: 0.9808 - val_loss: 0.0155 - val_accuracy: 1.0000\n",
            "Epoch 20/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.0704 - accuracy: 0.9788 - val_loss: 0.0139 - val_accuracy: 1.0000\n",
            "Epoch 21/30\n",
            "18/18 [==============================] - 1s 32ms/step - loss: 0.0611 - accuracy: 0.9808 - val_loss: 0.0121 - val_accuracy: 1.0000\n",
            "Epoch 22/30\n",
            "18/18 [==============================] - 1s 30ms/step - loss: 0.0592 - accuracy: 0.9827 - val_loss: 0.0113 - val_accuracy: 1.0000\n",
            "Epoch 23/30\n",
            "18/18 [==============================] - 1s 31ms/step - loss: 0.0629 - accuracy: 0.9808 - val_loss: 0.0091 - val_accuracy: 1.0000\n",
            "Epoch 24/30\n",
            "18/18 [==============================] - 1s 54ms/step - loss: 0.0983 - accuracy: 0.9654 - val_loss: 0.0104 - val_accuracy: 1.0000\n",
            "Epoch 25/30\n",
            "18/18 [==============================] - 1s 53ms/step - loss: 0.0505 - accuracy: 0.9827 - val_loss: 0.0100 - val_accuracy: 1.0000\n",
            "Epoch 26/30\n",
            "18/18 [==============================] - 1s 59ms/step - loss: 0.0628 - accuracy: 0.9827 - val_loss: 0.0128 - val_accuracy: 1.0000\n",
            "Epoch 27/30\n",
            "18/18 [==============================] - 1s 63ms/step - loss: 0.0472 - accuracy: 0.9865 - val_loss: 0.0078 - val_accuracy: 1.0000\n",
            "Epoch 28/30\n",
            "18/18 [==============================] - 1s 53ms/step - loss: 0.0526 - accuracy: 0.9827 - val_loss: 0.0086 - val_accuracy: 1.0000\n",
            "Epoch 29/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.0502 - accuracy: 0.9808 - val_loss: 0.0099 - val_accuracy: 1.0000\n",
            "Epoch 30/30\n",
            "18/18 [==============================] - 1s 33ms/step - loss: 0.0429 - accuracy: 0.9865 - val_loss: 0.0121 - val_accuracy: 1.0000\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_5_FWSMKIn-C"
      },
      "source": [
        "lstm.save(\"LSTMmodel.h5\")\n",
        "lst=models.load_model('LSTMmodel.h5')"
      ],
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "if1FZNAvtw6h"
      },
      "source": [
        "##Model testing and evaluation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xws0Cu8Yo1h9",
        "outputId": "aaa9b86c-bedd-46a0-b324-42ad795b514b"
      },
      "source": [
        "lstm.evaluate(X_test,Y_test)"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 17ms/step - loss: 0.0165 - accuracy: 0.9908\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.01653706096112728, 0.9908257126808167]"
            ]
          },
          "metadata": {},
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cEU7dNN7upVW",
        "outputId": "1d6dc059-ae4e-4b89-9b03-407d1cb92848"
      },
      "source": [
        "y_prediction1=lstm.predict(X_test)\n",
        "y_prediction1 = [np.argmax(element)for element in y_prediction1 ]\n",
        "\n",
        "print(classification_report(Y_test,y_prediction1))\n",
        "print(confusion_matrix(Y_test , y_prediction1))\n",
        "\n",
        "roc = roc_auc_score(Y_test, lstm.predict(X_test), multi_class ='ovr')\n",
        "kappa = cohen_kappa_score(Y_test, lstm.predict(X_test).argmax(axis=1))\n",
        "\n",
        "print(\"Roc score\", roc)\n",
        "print(\"Kappa score\", kappa)\n"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 11ms/step\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00        19\n",
            "           1       1.00      1.00      1.00        17\n",
            "           2       1.00      1.00      1.00        18\n",
            "           3       1.00      1.00      1.00        16\n",
            "           4       0.89      1.00      0.94         8\n",
            "           5       1.00      0.88      0.93         8\n",
            "           6       1.00      1.00      1.00        23\n",
            "\n",
            "    accuracy                           0.99       109\n",
            "   macro avg       0.98      0.98      0.98       109\n",
            "weighted avg       0.99      0.99      0.99       109\n",
            "\n",
            "[[19  0  0  0  0  0  0]\n",
            " [ 0 17  0  0  0  0  0]\n",
            " [ 0  0 18  0  0  0  0]\n",
            " [ 0  0  0 16  0  0  0]\n",
            " [ 0  0  0  0  8  0  0]\n",
            " [ 0  0  0  0  1  7  0]\n",
            " [ 0  0  0  0  0  0 23]]\n",
            "4/4 [==============================] - 0s 19ms/step\n",
            "4/4 [==============================] - 0s 13ms/step\n",
            "Roc score 1.0\n",
            "Kappa score 0.9890934560736442\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "djxoVGX-2PmF"
      },
      "source": [
        "#Example: pass a video to the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HAN1FEqxuEQE",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "478d85e8-cf52-4f4b-d299-1548decfa118"
      },
      "source": [
        "s='/content/drive/MyDrive/subject_6/like_04.mp4'\n",
        "cap=cv2.VideoCapture(s)\n",
        "i=0\n",
        "nb=0\n",
        "listeRL=[]\n",
        "while cap.isOpened():\n",
        "\n",
        "  imageId=cap.get(1)\n",
        "  ret,image=cap.read()\n",
        "  if ret !=True:\n",
        "    break\n",
        "\n",
        "  if imageId != 0:\n",
        "    if imageId % 4 == 0:\n",
        "      nb=nb+1\n",
        "      namefile=str(nb)+'.jpg'\n",
        "      cv2.imwrite(namefile,image)\n",
        "      image=imread(namefile)\n",
        "\n",
        "      if image.ndim==3:\n",
        "        image = cv2.GaussianBlur(image, (5,5) , 100)\n",
        "        red , green , blue = image[:,:,0] ,image[:,:,1] , image[:,:,2]\n",
        "\n",
        "\n",
        "        image_hsv = rgb2hsv(image)\n",
        "        hue = image_hsv[:,:,0]\n",
        "        sat = image_hsv[:,:,1]\n",
        "\n",
        "        image_ycbcr = rgb2ycbcr(image)\n",
        "        y = image_ycbcr[:,:,0]\n",
        "        cb = image_ycbcr[:,:,1]\n",
        "        cr = image_ycbcr[:,:,2]\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "        mask = (red > 95) & (green > 40) & (blue > 20) & (red > green) & (red > blue) & (np.abs(red - green) > 15) & (cr > 135) & (cb > 85) & (y > 80) & (cr <= (1.5862*cb)+20) & (cr >= (0.3448*cb)+76.2069) & (cr >= (-4.5652*cb)+234.5652) & (cr <= (-1.15 * cb)+301.75) & (cr <= (-2.2857*cb)+432.85)\n",
        "        mask = mask.astype(np.uint8)\n",
        "        mask[mask == 1] = 255\n",
        "        np.sum(mask == 255)\n",
        "\n",
        "\n",
        "        kernel = np.ones((5,5), np.uint8)\n",
        "        mask1 = cv2.erode(mask , kernel , iterations=2)\n",
        "\n",
        "        kernel = np.ones((5,5), np.uint8)\n",
        "        image = cv2.dilate(mask1 , kernel , iterations=4)\n",
        "        #image=cv2.cvtColor(image,cv2.COLOR_GRAY2RGB)\n",
        "        image=resize(image,(128,128))\n",
        "\n",
        "        listeRL.append(image)\n",
        "\n",
        "listeRL.pop()\n"
      ],
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       ...,\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.]])"
            ]
          },
          "metadata": {},
          "execution_count": 71
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HxHtoaAK2o5P"
      },
      "source": [
        "Mettre les frames en block"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ye5t9NlZM4P3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "68e757c4-1bbe-4a76-cedb-1cb99117668b"
      },
      "source": [
        "listeR=[]\n",
        "listeR=np.array(listeRL).reshape(-1,128,128,1)\n",
        "\n",
        "listeRR=[]\n",
        "listeRR=listeR/255\n",
        "\n",
        "listeR=[]\n",
        "listeR=modelfeatures.predict(listeRR)\n",
        "\n",
        "listo=[]\n",
        "lis=[]\n",
        "i=0\n",
        "for ph in listeR:\n",
        "  i=i+1\n",
        "  lis.append(ph)\n",
        "  if i==6:\n",
        "    listo.append(lis)\n",
        "    i=0\n",
        "    lis=[]\n",
        "\n",
        "listeR=np.array(listo).reshape(-1,6,150)\n",
        "print(listeR.shape)"
      ],
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 372ms/step\n",
            "(1, 6, 150)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZdYtRn0juELN",
        "outputId": "d6ba8a42-69d3-4e88-8a13-a86e47944e4f"
      },
      "source": [
        "y_pred=lstm.predict(listeR)\n",
        "print(\"probabilities list\", y_pred)\n",
        "\n",
        "ycla = [np.argmax(element)for element in y_pred ]\n",
        "i=0\n",
        "for i in ycla :\n",
        "  for j in label2int:\n",
        "    if label2int[j] == i:\n",
        "      print(j)"
      ],
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 60ms/step\n",
            "probabilities list [[2.0159056e-04 7.9890341e-01 1.9421242e-01 5.7913666e-04 4.2693717e-03\n",
            "  1.6389845e-03 1.9503101e-04]]\n",
            "like\n"
          ]
        }
      ]
    }
  ]
}