Get Started
Home
Topics
Search
Library
Research questionHow can robots recognize social gestures in real time despite occlusion, viewpoint changes, and noisy keypoints?Social cues such as invitations, refusals, and unavailability are not always explicit commands, while onboard observations can lose hands or arms as people move through changing viewpoints. Recognition must operate with noisy, unstable keypoints under robot latency constraints and accommodate newly emerging gesture classes.
AI
Computer Vision
Machine Learning
Robotics
Small / On-device Models
Latest papersRecent research connected to this question, newest first.SocioGesture: Real-Time and Adaptive Social Gesture Perception for Human-Robot InteractionThe source describes a body-and-hand skeleton-based perception system using dual streams for body motion and hand articulation. Evidence covers mixed indoor-outdoor HRI data, held-out-subject recognition, structured joint occlusion, real-time execution on a robot-mounted edge device, and deployment-time offline labeling of uncertain segments for vocabulary expansion.research paper · Sep 3, 2026
Related questions
How can an accessible humanoid robot integrate multimodal AI for real-world human interaction?How can real-time social robots reason over long-term context and decide when to act without disrupting fluent multimodal interaction?How can robotic manipulation policies combine vision, language, and touch for reliable contact-rich tasks under occlusion?How can dual-arm robots recover occluded object surfaces from calibrated, synchronized multi-view observations?