Fine Grained Classification and Image Denoising using CNNs

Carried out an extensive literature survey on various convolutional neural network architectures within 10M parameters to classify images in the CUB dataset
Performed hyperparameter tuning of learning rate, batch size, and number of epochs to improve accuracy from 54.69% to 73.92%
Designed and modified a network based on CBDNet to deblur images corrupted using Gaussian blur filters of different sizes and extent of blurring
Performed hyperparameter tuning to improve the peak signal-to-noise ratio from 26.68 to 28.42