Difference between revisions of "Occlusions"
| Line 7: | Line 7: | ||
<youtube>kBFMsY5ZP0o</youtube> | <youtube>kBFMsY5ZP0o</youtube> | ||
<youtube>_l5Pi7XcNOo</youtube> | <youtube>_l5Pi7XcNOo</youtube> | ||
| + | <youtube>4qGaARE8ob4</youtube> | ||
| + | <youtube>9oR_Ba0YOR8</youtube> | ||
Revision as of 22:52, 17 October 2018
- Conditional Adversarial Architecture (CAA) AI Sees Through Walls
Occlusion is a fundamental problem in human pose estimation and many other vision tasks. Instead of hallucinating missing body parts based on visible ones, we demonstrate a solution that leverages radio signals to accurately track the 2D human pose through walls and obstructions. Through-Wall Human Pose Estimation Using Radio Signals | Mingmin Zhao, Tianhong Li, Mohammad Abu, Alsheikh Yonglong, Tian Hang Zhao, Antonio Torralba, Dina Katabi - MIT CSAIL