New framework Multi-HMR 2 enhances human detection and 3D localization

Researchers have introduced Multi-HMR 2, a new framework designed for multi-person human detection, mesh recovery, and tracking within a camera-centric coordinate system. Unlike previous methods that focused on pelvis-centered recovery, this approach aims for accurate metric 3D localization, crucial for applications like human-robot interaction. The system leverages a DETR-based architecture and incorporates memory features from SAM2 to enable tracking without video supervision, achieving state-of-the-art results in detection accuracy and 3D localization. AI

RANK_REASON The cluster contains an academic paper detailing a new research framework for computer vision tasks. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• Multi-HMR 2
• Philippe Weinzaepfel
• SAM2

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →