My interests revolve around computer vision, deep learning and computer
graphics. I further wish to explore computer vision in 3D, and the recent developments in Neural Radiance Fields (NeRFs) really fascinate me. Interdisciplinary fields such as Computational Neuroscience and Computational Biology are also something I am keen to delve into. I am actively looking for research work related to computer vision,
computer graphics and computational neuroscience. if you would like to collaborate please drop an email!
I've worked as a research intern at TCS Research and as an undergraduate researcher at APPCAIR, our institute AI research lab. I have also worked as a research intern at
CEERI where I worked on digital restoration of ancient Indian wall paintings using deep learning techniques. We built an end to end pipeline including dataset preparation, damage detection and inpainting. Our research paper on the same has been accepted at International Conference for Computer Vision and Image Processing 2023.
Among other interests, I play Ultimate Frisbee and was the vice captain of the institute team for the year 2023-24.
I was also an active member of the Student Alumni
Relations Cell and Nirmaan (a social welfare club).
Feel free to check out my
CV
and drop me an
e-mail
if you want to chat with me!
Worked under the supervision of Pavan Kumar Chittimalli,
on the extraction of information and relations from large Entity-Relation (ER) images using deep learning techniques. My research spanned a diverse set of approaches, including heuristic-based image processing, object detection frameworks, and cutting-edge language models with prompt engineering aimming to optimize the accuracy and efficiency of information extraction in complex ER visualizations.
Student researcher
| APPCAIR & University of New South Wales (Australia)
Jan '24 - May '24
Worked under the supervision of Prof. Ashwin Srinivasan,
Prof. Erik Meijering,
Prof. Tanmay Verlekar,
Prof. Manik Gupta,
on segmentation of real time endoscopic footage (Hyper Kvasir dataset) and classification of skin diseases (HAM1000 dataset) using Deep Learning. Our main focus was to have computationally efficient models and deploy them on edge devices.
Remote Research Intern | North Eastern Space Applications Centre
Aug '23 - Nov '23
Working under the supervision of Ritu Anilkumar on glacier detection from satellite images of Sikkim and North Eastern India obtained from Copernicus satellite.
Machine Learning Research Intern
| CEERI Pilani
May '23 - Dec '23
Worked under under the supervision of Dr. Dhiraj Sangwan
on digital restoration of ancient Indian wall paintings using deep learning techniques. Devised an end to end pipeline, including creating a dataset of artificially damaged wall paintings, segmenting the damaged regions in the paintings and using inpainting techniques to restore them.
Hardaat Singh Baath, Soham Shinde, Jinam Keniya, Priyanshu Ranjan Mishra,
Anil Saini, Dhiraj Sangwan - "Damage Segmentation and Restoration of Ancient
Wall Paintings for Preserving Cultural Heritage", Accepted
at International Conference on Computer Vision and Image Processing, 2023
Paper implementation of NeurIPS Spotlight Rank-N-Contrast
Reproduced the results of the paper on the AgeDB dataset. Along with the official reproduction, tried out this loss on a Graph Regression Task, the ESOL dataset.
Damage restoration of ancient Indian wall paintings using Deep
Learning
Worked on identifying damaged regions from damaged ancient Indian murals and
using image inpainting techniques to digitally restore them. We initially
created an exhaustive dataset by synthetically damaging clean images of wall
paintings using binary masks and textures, followed by using segmentation
models for damage segmentation and GAN architecture for image inpainting.
Working on finding the most optimum locations for placing TSVs on a 3D (NoC) Network-on-Chip using reinforcement learning and various optimization techniques.
This template is a modification to Jon Barron's website
and a fork of Rishab Khincha's website.
Find the source code to my website here.