Ashwath Shetty

I'm a Research Assistant at the Computer Vision and Machine Learning Department, at the Max Planck Insitute working with Jonas Fischer on understanding the effect of sparsity on network intepretibility .

Before that, I did my Masters Thesis student at the Visual Computing and AI Department, at the Max Planck Insitute in Saarbrücken. I did my thesis working on Real time photorealistic telepresence with Marc Habermann and Christian Theobalt. Previosly I worked as Research Assistant under Avinash Sharma on 3D Performance Capture at the CVIT Lab in IIIT Hyderabad. I got my Bachelor degree from beautiful MIT Manipal Karnataka, and did my bachelors thesis at IIT Bombay under Ganesh Ramakrishnan.

Hobbies and free time activities:Running, Meditation, Hiking, Martial Arts, Wind instruments (Bamboo Flute and Harmonica), Irish Dancing, Chess (I'm just average in all of them :))

Email  /  CV  /  DBLP  /  Github

Research

I'm primarily interested in computer vision, machine learning, optimization, geometry processing and the intersection of these fields. Most of my previous research has covered diverse aspects of modelling humans in motion in a photorealistic and efficient way, however I am eager to learn more about robust representation learning, interpretibility, and neuroscience. If you have any projects where you think I could be useful or could learn from, feel free to reach out.

Selected Experience (See CV for more)

Research Assistant at the CVML Department, Max Planck Insitute for Informatics ( supervised by Jonas Fischer) (March 2024 -Present)

• Sparse Compositional interpretibility Understanding the link between sparsity and interpretability, and using a sparse network to read off a networks decision making process

Research Assistant at the VCAI Department, Max Planck Insitute for Informatics ( supervised by Marc Habermann and Christian Theobalt) (July 2022-Jan 2024)

• Holoported Characters Built an animatable human representation, that takes an sparse images and 3D pose, and can render humans in arbitrary viewpoints at 4K resolution (Selected as one of the inaugaral VIA centre projects)

Research Assistant at the Centre for Visual Information Technology, IIIT Hyderabad ( supervised by Avinash Sharma) (July 2020-Jan 2022)

• 3D SMPL+D Registration Setup a multi stage pipeline for fitting SMPL followed by SMPL+D to a 3D human scan. (Sample Results)
• Face Deocclusion Research on Face Deocclusion and Reconstruction, from videos
• Some other cool projects Data generation pipeline from stylgan and styleflow (Sample Results), Event Camera based SMPL Tracking

Research Intern at DreamVu ( supervised by Avinash Sharma and Anoop Namboodiri) (July 2022-Present)

• 3D Dance Capture Worked on TSDF fusion mesh details, with deep learning models for faces (Sample Results)

Online Courses and Summer Schools

I enjoy learning, so I do spend a lot of my time doing online courses in random and sometimes relevant topics.

Certified: (Link to Certificates) Algorithms (Stanford), Probabilistic Graphical Models (Stanford), Deep Learning (deeplearning.ai), Tensorflow in Practice (deeplearning.ai), Machine Learning(Stanford), Advanced Machine Learning (HSE), Reinforcement Learning (UAlberta), Programming Languages (UWash)

Online Audit: Calculus (MIT), Math for CS (MIT), Linear Algebra (MIT), CNNs for Vision (Stanford), NLP with Deep Learning (Stanford), Learning from Data (Caltech), Interactive Computer Graphics (Udacity), Computer Vision (Udacity\&Gtech), Practical DL for coders (fast.ai), Deep Learning (NYU)

Summer Schools: Easter European Machine Learning, ACM Graph Theory and Graph Algorithms

Selected Projects (Refer to CV for more)

Feel free to steal this website's source code. Do not scrape the HTML from this page itself, as it includes analytics tags that you do not want on your own website — use the github code instead. Also, consider using Leonid Keselman's Jekyll fork of this page.