Research code for CVPR 2021 paper “End-to-End Human Pose and Mesh Reconstruction with Transformers”
Image and video have become the language people use to communicate on the Internet. Multimedia content connects people and appeals to the young. This project aims at deep image and video transformation to generate high-quality…
We have been developing SOTA technologies and industry-leading product solutions for following scenarios: (1) Universal OCR to detect and recognize any text in image/PDF; (2) Universal math OCR to detect and recognize any math expression…
Modern work increasingly relies on online collaboration with real-time communications (RTC). Our research aims to provide real-time, intelligent, and immersive media experiences, with a long-term vision of advancing multimedia technologies in a manner that shapes…
This project is part of the multi-sense efforts within the people centric strategy of Microsoft. It addresses a number of vertical domains for AI by developing effective human-centric spatial understanding technologies to extract insights from…
