Publications
(*equal contribution, ✉ corresponding author)
Email: medhi.moushumi@gmail.com
Moushumi Medhi
PhD Candidate @ IIT Kharagpur
Hello! I am a Ph.D student at Indian Institute of Technology (IIT) Kharagpur, advised by Prof. Rajiv Ranjan Sahay. Previously, I worked as a Research Consultant in the Department of Electrical Engineering at IIT Kharagpur with Professor Sahay, on an industry-sponsored project funded by and delivered to Altair Engineering India Pvt. Ltd., Bangalore. Before that, I received my M.Tech. in Electronics and Communication Engineering from Tezpur Central University, where I worked on a government-funded project under the Council of Scientific and Industrial Research (CSIR), carried out at the CSIR–Central Electronics Engineering Research Institute (CSIR-CEERI), Pilani. As a master's dissertation intern at CSIR–CEERI, I was supervised by Dr. Ing. Jagdish Lal Raheja. I hold a B.Tech. in Electronics and Telecommunication Engineering from Assam Engineering College.
My research interests broadly lie in Machine Learning and Computer Vision, with a focus on generative AI and 3D vision. During my PhD, my thesis work has centered on developing and optimizing lightweight generative models to improve speed and efficiency of ill-posed 2D-3D vision problems (e.g., depth restoration, depth estimation) under limited data and constrained compute resources. Beyond my thesis, my research has explored the integration of physical image formation models with learning-based approaches. This includes work on depth from defocus, geometry-aware scene reconstruction from light field images, and generative representations for realistic view synthesis and restoration. I have also engaged with emerging topics at the intersection of vision and language, augmented reality, diffusion-based generative modeling, and classification and pattern recognition-based problems.