Universal Feature Representation

Universal compressed image features optimized for multi-task neural networks

Compression for Machines

Framework for Compression for machines for vision tasks: Object Detection, Semantic Segmentation, Classificaiton (Supervisor: Dr. Ivan Bajić)

Math2latex

Converting math equations to latex code using vision transformers

Diabetes Retinopathy Classification

Diabetes Retinopathy detection and classification using OCT and Fundus

SVM-IPM

Implementation of Soft margin Support Vector Machines (SVM) using Interior Point Methods: 1)Barrier method and 2)Primal-Dual method

OpenAI-Cartpole

Solving OpenAI Cartpole problem using Q-Learning and Deep Q-Networks