Hello! I am Bindita Chaudhuri. I am a 5th (final) year Ph.D. student in the Paul G. Allen School of Computer Science and Engineering at University of Washington. I work in the Graphics and Imaging Laboratory under the guidance of Prof. Linda Shapiro. I also collaborate with Alex Colburn, Barbara Mones and Gary Faigin.
My research interests are in the fields of computer vision, machine learning, computer graphics and computational photography. I have worked on developing real-time deep learning frameworks to automatically transfer human facial motion from 2D images to 3D characters for immersive VR/AR/MR storytelling and communication. At present, I am working on photorealistic human avatar generation and 2D-to-3D body motion retargeting.
Previously, I completed my Masters in Electrical Engineering from Indian Institute of Technology Bombay under the guidance of Prof. Subhasis Chaudhuri in 2016 and received my Bachelors degree in Electronics and Telecommunication Engineering from Jadavpur University under the guidance of Prof. Iti Saha Misra in 2014.
Virtual Humans team, Facebook Reality Labs / Oculus Research (Sausalito, CA)
Photorealistic texture synthesis for 3D humans
AI Perception and Mixed Reality Platform team, Microsoft Cloud & AI (Redmond, WA)
Personalized face modeling for high-fidelity 3D reconstruction and improved 3D tracking and retargeting from 2D images.
Visual Intelligence group, Microsoft Research (Redmond, WA)
Designed a multi-task deep learning framework to transfer performance of single or multiple human faces captured with a 2D camera to 3D animated characters in real time.
Computational Imaging Lab, Intel Labs (Santa Clara, CA)
Designed convolutional neural networks to perform optical flow prediction and image super-resolution for efficient view synthesis from high-definition multi-camera array images.
US patent granted!
Department of Information Engineering and Computer Science, University of Trento (Trento, Italy)
Worked with Prof. Lorenzo Bruzzone and Prof. Begum Demir on novel unsupervised and semi-supervised approaches to content-based retrieval of remote sensing images using an inexact graph matching strategy.
University of Washington (Seattle, WA)
Courses taught: Computer Vision, Artificial Intelligence, Algorithms, Compilers