Person re-identification system speeds up multi-camera searches
A Generative AI Integrated Multimodal Framework for Low-Latency Multi-Camera Person Re-Identification
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Finding the same person across multiple camera views is hard because of changes in angle, lighting, and partial occlusions. The authors present a system that uses different types of clues—from the overall look of the person to detailed descriptions generated by AI models and sometimes facial features—to quickly identify people. To keep the process fast, their system first tries simple clues and only uses more complex information when the initial results are unclear. They showed that this method works well and saves computing time on standard public datasets.
What this means in practice
- •For security system developers: Deploy multi-camera person search that reduces processing time by rejecting ambiguous cases early and applying detailed analysis only when needed.
- •For video analytics platform teams: Design adaptive person tracking modules that combine visual and AI-generated semantic clues to maintain accuracy with lower computational costs.
Authors
Leon Fernando, C Dombawala, P. Hettigoda, Vanodhya G. Warnasooriya, Ishara Neranjana, Rashmika Nawaratne
Abstract
Person re-identification (ReID) is essential for multi-camera surveillance and tracking, yet remains difficult due to viewpoint and illumination changes, occlusion, background clutter, and low resolution imagery. We propose a generative AI integrated multimodal ReID framework designed explicitly for robustness under missing cues and low latency deployment. The key idea is a cost aware early-exit cascade that prioritizes inexpensive, high confidence evidence and only triggers expensive modalities for ambiguous cases. Our system integrates (i) global visual embeddings from segmented person regions, (ii) automatically generated fine grained semantic attribute descriptions generated by vision-language models (VLMs), and (iii) optional facial embeddings when face observations are reliable. To optimize the balance between accuracy and latency, we use a cost aware early-exit cascade instead of fusing all modalities. Specifically, we first inspect the top-k retrieval results to determine whether the query is unambiguous. If the best match is clearly separated from the remaining candidates, we stop early and return the result to minimize latency; in ambiguous cases, we keep multiple hypotheses and invoke additional modalities (face/semantic) with adaptive reliability weighting to refine the decision. We report person re-identification performance using mAP and Rank-1 accuracy on the Market-1501 and DukeMTMC-reID benchmarks. The proposed adaptive early-exit cascade resolves 60.7% of DukeMTMC-reID queries and 68.4% of Market-1501 queries without invoking semantic reasoning, reducing computational overhead while maintaining competitive retrieval performance.