Object-Detection

This Repository contains all the files related to Object Detection model using YOLO algorithm.

Important points related to model:

• This a Deep CNN model that uses YOLO(You Only Look Once) algorithm for identifying objects.
• It includes a single scan of the image to detect the objects.
• There are various anchor boxes that are used to bound the objects and prediction is done for each anchor box.
• The model uses transfer learning to get the pre-trained parameters and was then fine-tuned with cross validation set to improve the test-set accuracy.
• Model uses Non-Max Suppression and Intersection-over-Union(IoU) as a parameter to filter multiple anchor boxes.
• It was developed on Keras Framework.