A5044972952- Hand Gesture Recognition Systems
- Video Surveillance and Tracking Methods
- Human Pose and Action Recognition
- Tactile and Sensory Interactions
- Hearing Impairment and Communication
- Artificial Intelligence in Healthcare and Education
- Ethics and Social Impacts of AI
- Infrared Thermography in Medicine
- Advanced Vision and Imaging
- AI in Service Interactions
Microsoft Research (India)
2023
Indian Institute of Science Bangalore
2022
The advances in monocular 3D human pose estimation are dominated by supervised techniques that require large-scale 2D/3D annotations. Such methods often behave erratically the absence of any provision to discard unfamiliar out-of-distribution data. To this end, we cast learning as an unsupervised domain adaptation problem. We introduce MRP-Net <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> Project page:...
Available 3D human pose estimation approaches leverage different forms of strong (2D/3D pose) or weak (multi-view depth) paired supervision. Barring synthetic in-studio domains, acquiring such supervision for each new target environment is highly inconvenient. To this end, we cast learning as a self-supervised adaptation problem that aims to transfer the task knowledge from labeled source domain completely unpaired target. We propose infer image-to-pose via two explicit mappings viz....
The Deaf or Hard-of-Hearing (DHH) community constitutes over 430 million people globally, with about 70 of them using sign language as their primary means communication. Learning poses significant challenges, and researchers have explored the potential digital games educational tools for teaching language. However, existing are limited to a narrow range words feature only single-player modes. Consequently, our objective was design language-based game that not facilitates learning but also...
Large Language Models (LLMs) have revolutionized programming and software engineering. AI assistants such as GitHub Copilot X enable conversational programming, narrowing the gap between human intent code generation. However, prior literature has identified a key challenge--there is user's mental model of system's understanding after sequence natural language utterances, actual understanding. To address this, we introduce Programming with Representations (PwR), an approach that uses...
Conventional domain adaptation algorithms aim to achieve better generalization by aligning only the task-discriminative causal factors between a source and target domain. However, we find that retaining spurious correlation non-causal plays vital role in bridging gap improving adaptation. Therefore, propose build framework disentangles supports factor alignment first. We also investigate strong shape bias of vision transformers, coupled with its multi-head attention, make it suitable...
The Deaf or Hard-of-Hearing (DHH) community constitutes over 430 million people globally, with about 70 of them using sign language as their primary means communication. India has around 63 DHH individuals. in faces several challenges, particularly learning and English, due to delayed diagnosis, stigma, oralism, a diversity languages. Digital games for spoken have gained popularity advantages traditional methods, such enhanced engagement socialization, driving increased research adoption...
The advances in monocular 3D human pose estimation are dominated by supervised techniques that require large-scale 2D/3D annotations. Such methods often behave erratically the absence of any provision to discard unfamiliar out-of-distribution data. To this end, we cast learning as an unsupervised domain adaptation problem. We introduce MRP-Net constitutes a common deep network backbone with two output heads subscribing diverse configurations; a) model-free joint localization and b)...