Sing Bing Kang
Sing Bing Kang is a computer vision researcher whose work spans robot instruction, image-based rendering, and indoor mapping. He earned a robotics PhD at Carnegie Mellon in 1994 and, by 2012, was a principal researcher at Microsoft Research, where he described computer vision projects for graphics. His projects included Holoportation and research on privacy risks in 3-D maps; at Zillow, he coauthored an indoor panorama dataset and was named as an inventor on a mobile panorama patent. He chaired ACCV and CVPR and became an IEEE Fellow. Since March 2025, he has been a Distinguished Scientist, AI at Zillow Group.
- Now at
- Zillow Group
- Works
- 301
- Citations
- 17,606
- h-index
- 71
Publication numbers from OpenAlex, as of 15 October 2026.
Timeline
-
1994
Robot instruction by demonstration
He completes a robotics PhD at Carnegie Mellon in 1994 with the thesis Robot Instruction by Human Demonstration. The thesis describes a system that observes a person perform a task, recognizes the grasp, and maps it onto a manipulator.
-
1997
A robotics paper award
His paper on mapping human hand motion to robot grasps receives the King-Sun Fu Memorial Best Transaction Paper Award in 1997.
-
2007–2012
Conference chairs and IEEE fellowship
He chairs ACCV in 2007 and CVPR in 2009, and becomes an IEEE Fellow in 2012.
-
2012
Computer vision for graphics
In 2012, he is a principal researcher at Microsoft Research and describes projects that use computer vision to animate Chinese paintings, generate free-viewpoint video, create 3-D models of plants and trees, and enhance photographs. He coauthors Image-Based Rendering and Image-Based Modeling of Plants and Trees.
-
2016
Holoportation in real time
In 2016, he coauthors Holoportation, a 3-D capture system that reconstructs, compresses, and transmits 3-D models of people in real time.
-
2019–2021
Privacy and indoor mapping
In 2019, he coauthors research showing that 3-D point clouds can retain enough information to reconstruct detailed images of mapped scenes after source images are discarded. He coauthors the 2021 CVPR paper on the Zillow Indoor Dataset, which contains 67,448 panoramas from 1,575 unfurnished homes, annotated with 3-D room layouts, 2-D boxes for windows and doors, and floor plans.
-
2021–2022
Mobile panorama generation
In 2021, he is named as an inventor on a Zillow patent for automated generation of panorama images on mobile devices for building locations. The patent publishes in 2022.
-
Since March 2025
A part-time Zillow role
Since March 2025, he is a Distinguished Scientist, AI at Zillow Group. His role includes on-device panorama stitching for Zillow's 3D Home Tour.
Sources and credits
Photo: We have no published photo of Sing Bing Kang yet. Send us a free-licensed photo.
- Robot Instruction by Human Demonstration
- Sing Bing Kang -- Talk Abstract
- Using Computer Vision for Graphics - Microsoft Research
- Holoportation: Virtual 3D Teleportation in Real-Time - Microsoft Research
- Envisioning privacy preserving image-based localization for augmented reality - Microsoft Research
- Zillow Indoor Dataset (ZInD) - GitHub
- Automated generation on mobile devices of panorama images for building locations and subsequent use - Google Patents
- Sing Bing Kang | alphaXiv