In a group of 3, we worked on 3 projects
-
Project 1: Classification( hotdog/not hotdog) and object detection using CNN and sliding window to localize house numbers using SVHN dataset
-
Project 2: segmentation implementing CNN (SegNet and Unet architectures) using LIDC-IDRI dataset 3:
-
Project 3: implementation of a fully connected neural network that generates numbers/clothes from MNIST/ Fashion MNIST dataset and a CycleGAN that can convert images of horses to zebras, and vice versa.