New MHRGait method advances gait recognition using articulated controls

Researchers have introduced MHRGait, a novel approach to gait recognition that utilizes Momentum Human Rig (MHR) pose parameters. This method represents gait using 184 semantically organized body and hand parameters derived from monocular video, modeling their intra-frame coordination and temporal evolution. Experiments on benchmarks like CCPG and SUSTech1K demonstrate that MHRGait achieves superior performance among model-based methods and exhibits effective cross-dataset transferability with a relatively small network size. An extension, MHRGait++, further enhances silhouette-based recognition by fusing MHR pose information, offering a favorable accuracy-efficiency trade-off. AI

IMPACT Introduces a novel representation for gait recognition, potentially improving accuracy and efficiency in surveillance and human-computer interaction applications.

RANK_REASON Research paper detailing a new method for gait recognition. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• FAM120B
• Hugging Face
• MHRGait
• Momentum Human Rig (MHR) pose
• SUSTech1K

• paper
• other

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