# Unsupervised Video Person Re-identification via Noise and Hard frame Aware Clustering

arXiv: https://arxiv.org/abs/2106.05441 · HTML: https://mlsanity.com/paper/2106.05441/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 4 of its 14 references that are in the map.
- First posted on arXiv 10 Jun 2021.

## Builds on (most cited first, 4 of 4)

- [Joint Detection and Identification Feature Learning for Person Search](https://mlsanity.com/paper/1604.01850/) (2016)
- [Unsupervised Tracklet Person Re-Identification](https://mlsanity.com/paper/1903.00535/) (2019)
- [Dynamic Label Graph Matching for Unsupervised Video Re-Identification](https://mlsanity.com/paper/1709.09297/) (2017)
- [Deep Association Learning for Unsupervised Video Person Re-identification](https://mlsanity.com/paper/1808.07301/) (2018)

## Built on by (most cited first, 3 of 3)

- [Exploiting Robust Unsupervised Video Person Re-identification](https://mlsanity.com/paper/2111.05170/) (2021)
- [Unleashing the Potential of Tracklets for Unsupervised Video Person Re-Identification](https://mlsanity.com/paper/2406.14261/) (2024)
- [Temporal Prototyping and Hierarchical Alignment for Unsupervised Video-based Visible-Infrared Person Re-Identification](https://mlsanity.com/paper/2604.21324/) (2026)

## Abstract (arXiv)

Unsupervised video-based person re-identification (re-ID) methods extract richer features from video tracklets than image-based ones. The state-of-the-art methods utilize clustering to obtain pseudo-labels and train the models iteratively. However, they underestimate the influence of two kinds of frames in the tracklet: 1) noise frames caused by detection errors or heavy occlusions exist in the tracklet, which may be allocated with unreliable labels during clustering; 2) the tracklet also contains hard frames caused by pose changes or partial occlusions, which are difficult to distinguish but informative. This paper proposes a Noise and Hard frame Aware Clustering (NHAC) method. NHAC consists of a graph trimming module and a node re-sampling module. The graph trimming module obtains stable graphs by removing noise frame nodes to improve the clustering accuracy. The node re-sampling module enhances the training of hard frame nodes to learn rich tracklet information. Experiments conducted on two video-based datasets demonstrate the effectiveness of the proposed NHAC under the unsupervised re-ID setting.