A CNN used to classify 7000 images into 8 classes. Used image augmentation technique to add another 7000 images in training set. Used vgg16 as base model, applied transfer learning to get better accuracy. Final classification accuracy is apporximately 75%.