Computer_Vision_Deep_Learning

phase 1

• creating a simple CNN model for training on the MNIST dataset

phase 2

• classifying images in The Oxford-IIIT Pet Dataset with feature extraction from pre-trained popular models including res-net, mobile-net, and google-net, and giving the extracted features to the SVM model for classification features extracted from average pool layers are the best because they are at the end of each model

phase 3

• creating a simple cluster model based on the features extracted from average pool layers of popular models