Jiangmiao Pang
Jiangmiao Pang is a research scientist at Shanghai AI Laboratory, where he heads the Embodied AI Center in collaboration with Intern Robotics. He is best known for his work on humanoid robotics and embodied AI, including whole-body motion control, robotic manipulation, and simulation systems for embodied agents. He is a co-corresponding author on Imagine2Real and has contributed to widely used open-source computer vision toolkits including MMDetection and MMDetection3D, accumulating over 17,000 citations on Google Scholar.
“Imagine2Real: Towards Zero-shot Humanoid-Object Interaction via Video Generative Priors”
Source→“Real-world deployment relies on motion capture systems, which are prone to marker occlusion during close-contact tasks (e.g., pushing boxes). Future work could integrate self-tracking systems using on-device SLAM and multi-camera fusion.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.