Visualization

nobrainer/visualization/Guided-GradCAM.py

@@ -0,0 +1,31 @@
+import numpy as np
+from skimage.transform import resize
+import tensorflow as tf
+
+
+def Guided_GradCAM_3D(grad_model, ct_io, class_index):
+    # First outputs target convolution and output
+    input_ct_io = tf.expand_dims(ct_io, axis=-1)
+    input_ct_io = tf.expand_dims(input_ct_io, axis=0)
+    with tf.GradientTape() as tape:
+        conv_outputs, predictions = grad_model(input_ct_io)
+        loss = predictions[:, class_index]
+    # Extract filters and gradients
+    output = conv_outputs[0]
+    grads = tape.gradient(loss, conv_outputs)[0]
+    guided_grads = (
+        tf.cast(output > 0, "float32") * tf.cast(grads > 0, "float32") * grads
+    )
+
+    # Average gradients spatially
+    weights = tf.reduce_mean(guided_grads, axis=(0, 1, 2))
+    # Build a ponderated map of filters according to gradients importance
+    cam = np.ones(output.shape[0:3], dtype=np.float32)
+    for index, w in enumerate(weights):
+        cam += w * output[:, :, :, index]
+
+    capi = resize(cam, (ct_io.shape))
+    print(capi.shape)
+    capi = np.maximum(capi, 0)
+    heatmap = (capi - capi.min()) / (capi.max() - capi.min())
+    return heatmap