KeyRe-ID: Keypoint-Guided Person Re-Identification using Part-Aware Representation in Videos

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Jinseong, Song, Jeonghoon, Baek, Gyeongseon, Noh, Byeongjoon
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915394741075968
author Kim, Jinseong
Song, Jeonghoon
Baek, Gyeongseon
Noh, Byeongjoon
author_facet Kim, Jinseong
Song, Jeonghoon
Baek, Gyeongseon
Noh, Byeongjoon
contents We propose \textbf{KeyRe-ID}, a keypoint-guided video-based person re-identification framework consisting of global and local branches that leverage human keypoints for enhanced spatiotemporal representation learning. The global branch captures holistic identity semantics through Transformer-based temporal aggregation, while the local branch dynamically segments body regions based on keypoints to generate fine-grained, part-aware features. Extensive experiments on MARS and iLIDS-VID benchmarks demonstrate state-of-the-art performance, achieving 91.73\% mAP and 97.32\% Rank-1 accuracy on MARS, and 96.00\% Rank-1 and 100.0\% Rank-5 accuracy on iLIDS-VID. The code for this work will be publicly available on GitHub upon publication.
format Preprint
id arxiv_https___arxiv_org_abs_2507_07393
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle KeyRe-ID: Keypoint-Guided Person Re-Identification using Part-Aware Representation in Videos
Kim, Jinseong
Song, Jeonghoon
Baek, Gyeongseon
Noh, Byeongjoon
Computer Vision and Pattern Recognition
Artificial Intelligence
We propose \textbf{KeyRe-ID}, a keypoint-guided video-based person re-identification framework consisting of global and local branches that leverage human keypoints for enhanced spatiotemporal representation learning. The global branch captures holistic identity semantics through Transformer-based temporal aggregation, while the local branch dynamically segments body regions based on keypoints to generate fine-grained, part-aware features. Extensive experiments on MARS and iLIDS-VID benchmarks demonstrate state-of-the-art performance, achieving 91.73\% mAP and 97.32\% Rank-1 accuracy on MARS, and 96.00\% Rank-1 and 100.0\% Rank-5 accuracy on iLIDS-VID. The code for this work will be publicly available on GitHub upon publication.
title KeyRe-ID: Keypoint-Guided Person Re-Identification using Part-Aware Representation in Videos
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2507.07393