Social Interaction Modeling for Group Re-identification

ICLR 2026 Conference Submission18002 Authors

Keywords: Group Re-identification; Member Differentiation; Social Interaction; Attention Mechanism; Layout Variation

TL;DR: Greid

Abstract: Group Re-identification (G-ReID) focus on associating group images that contain the same members across different camera views. The key challenge is that identity differentiation and position differentiation in group topology structure changes are difficult to capture. According to the social psychology principles, we found that the core members are more likely to remain in the group with smaller position changes, and peripheral members are more likely to have significant position changes or even fade out of the group. To this end, we propose a novel social interaction modeling (SIM), which treats group as a social interaction field, explore more authentic and robustness group features through the member differentiation. The member differentiation contains identity and position differentiation. Our method constructs the social interaction calculation module (SICM) to capture the member differentiation in fields, and implements identity differentiation and position differentiation by the social prior attention mechanism (SPAM) and social layout variation module (SLVM), respectively. A large number of experiments have been conducted on three available datasets show that the proposed method SIM is effective, and outperforms all previous state-of-the-art methods, surpassing the baseline on Rank1/mAP by up to 8.6\%/9.6\% on DukeGroup, 3.7\%/2.7\% on RoadGroup and 2.5\%/2.9\% on CSG. Code will be available on github.

Primary Area: applications to computer vision, audio, language, and other modalities

Submission Number: 18002