Privacy-preserving person re-identification uses depth images and Transformers

Researchers have developed a novel person re-identification system that prioritizes privacy by utilizing depth images instead of traditional RGB data. This method employs a Transformer encoder to process temporal sequences, capturing dynamic movement patterns from both RGB and depth modalities. The Hungarian algorithm is used to optimize the matching process between different views, and batch hard triplet loss enhances feature learning. Evaluations on datasets like TVPR2 and BIWI RGBD-ID show that the depth-only approach achieves competitive performance while preserving user privacy. AI

RANK_REASON The cluster contains a research paper detailing a novel method for person re-identification. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• BIWI RGBD-ID
• Hungarian algorithm
• Raphaël Delécluse
• Transformer
• TVPR2

• paper
• other

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