# Rethinking Temporal Fusion for Video-based Person Re-identification on Semantic and Time Aspect

arXiv: https://arxiv.org/abs/1911.12512 · HTML: https://mlsanity.com/paper/1911.12512/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 28 Nov 2019.

## Built on by (most cited first, 8 of 8)

- [TIPCB: A Simple but Effective Part-based Convolutional Baseline for Text-based Person Search](https://mlsanity.com/paper/2105.11628/) (2021)
- [Spatial-Temporal Correlation and Topology Learning for Person Re-Identification in Videos](https://mlsanity.com/paper/2104.08241/) (2021)
- [Multi-direction and Multi-scale Pyramid in Transformer for Video-based Pedestrian Retrieval](https://mlsanity.com/paper/2202.06014/) (2022)
- [Activity-Biometrics: Person Identification from Daily Activities](https://mlsanity.com/paper/2403.17360/) (2024)
- [CoNAN: Conditional Neural Aggregation Network For Unconstrained Face Feature Fusion](https://mlsanity.com/paper/2307.10237/) (2023)
- [DisenQ: Disentangling Q-Former for Activity-Biometrics](https://mlsanity.com/paper/2507.07262/) (2025)
- [Multi Camera Connected Vision System with Multi View Analytics: A Comprehensive Survey](https://mlsanity.com/paper/2510.09731/) (2025)
- [KeyRe-ID: Keypoint-Guided Person Re-Identification using Part-Aware Representation in Videos](https://mlsanity.com/paper/2507.07393/) (2025)

## Abstract (arXiv)

Recently, the research interest of person re-identification (ReID) has gradually turned to video-based methods, which acquire a person representation by aggregating frame features of an entire video. However, existing video-based ReID methods do not consider the semantic difference brought by the outputs of different network stages, which potentially compromises the information richness of the person features. Furthermore, traditional methods ignore important relationship among frames, which causes information redundancy in fusion along the time axis. To address these issues, we propose a novel general temporal fusion framework to aggregate frame features on both semantic aspect and time aspect. As for the semantic aspect, a multi-stage fusion network is explored to fuse richer frame features at multiple semantic levels, which can effectively reduce the information loss caused by the traditional single-stage fusion. While, for the time axis, the existing intra-frame attention method is improved by adding a novel inter-frame attention module, which effectively reduces the information redundancy in temporal fusion by taking the relationship among frames into consideration. The experimental results show that our approach can effectively improve the video-based re-identification accuracy, achieving the state-of-the-art performance.