• Home
  • Search
  • Camera-Aware Differentiated Clustering With Focal Contrastive Learning for Unsupervised Vehicle Re-Identification
  • Cite Icon4
  • https://doi.org/10.1109/tcsvt.2024.3402109Copy DOI Icon

Camera-Aware Differentiated Clustering With Focal Contrastive Learning for Unsupervised Vehicle Re-Identification

  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Most existing research on vehicle re-identification (Re-ID) focuses on supervised methods, while unsupervised methods that can take advantage of massive unlabeled data are underexplored. Due to the similarity of tasks, unsupervised person Re-ID methods that employ clustering to generate pseudo labels for model training can achieve good performance on unsupervised vehicle Re-ID task. However, vehicle exhibit higher intra-ID compactness and inter-ID separability within camera than person, which has not been exploited to reduce pseudo label noise for unsupervised vehicle Re-ID. To address this issue, we propose a camera-aware differentiated clustering with focal contrastive learning (CDF) method for unsupervised vehicle Re-ID task. Unlike the conventional global clustering approach that adopts a uniform processing strategy for pseudo-label generation, a camera-aware differentiated clustering (CDC) approach is designed to reduce label noise. In CDC, the entire clustering process is divided into two stages: inter-camera and intra-camera clustering, and each stage adopts different clustering strategies that are carefully designed according to the differences in feature distribution within and across cameras. By considering the distribution of pseudo labels generated by CDC, a measure for calculating the reliability of inter-camera and intra-camera pseudo labels is further designed, and a focal contrastive learning loss is proposed to improve the model’s ID discrimination ability within and across cameras. Extensive experiments on VeRi-776 and VERI-Wild demonstrate the effectiveness of each designed component and the superiority of the CDF.

Similar Papers
  • Conference Article
  • Citations40

Enhancing Diversity in Teacher-Student Networks via Asymmetric branches for Unsupervised Person Re-identification

  • Jan 01, 2021
  • Hao Chen +2
  • PDF
  • Research Article
  • Citations4

Unsupervised Vehicle Re-Identification Based on Cross-Style Semi-Supervised Pre-Training and Feature Cross-Division

  • Jul 03, 2023
  • Electronics
  • Guowei Zhan +5
  • Research Article
  • Citations5

Pro-ReID: Producing reliable pseudo labels for unsupervised person re-identification

  • Aug 28, 2024
  • Image and Vision Computing
  • Haiming Sun +1
  • Research Article
  • Citations33

Mask-Aware Pseudo Label Denoising for Unsupervised Vehicle Re-Identification

  • Apr 01, 2023
  • IEEE Transactions on Intelligent Transportation Systems
  • Zefeng Lu +3
  • Research Article
  • Citations32

Diversity feature constraint based on heterogeneous data for unsupervised person re-identification

  • Feb 02, 2023
  • Information Processing & Management
  • Tongzhen Si +4
  • Research Article
  • Citations27

Relation-Preserving Feature Embedding for Unsupervised Person Re-Identification

  • Jan 01, 2024
  • IEEE Transactions on Multimedia
  • Xueping Wang +5
  • Research Article
  • Citations37

Viewpoint-Aware Progressive Clustering for Unsupervised Vehicle Re-Identification

  • Aug 01, 2022
  • IEEE Transactions on Intelligent Transportation Systems
  • Aihua Zheng +3
  • Research Article
  • Citations24

Pseudo Label Based on Multiple Clustering for Unsupervised Cross-Domain Person Re-Identification

  • Jan 01, 2020
  • IEEE Signal Processing Letters
  • Shuni Chen +2
  • Research Article
  • Citations18

Distribution-Guided Hierarchical Calibration Contrastive Network for Unsupervised Person Re-Identification

  • Aug 01, 2024
  • IEEE Transactions on Circuits and Systems for Video Technology
  • Yongxi Li +4
  • Research Article

Camera Style Alignment for Unsupervised Person Re-Identification

  • Jan 01, 2025
  • IEICE Transactions on Information and Systems
  • Jifeng Guo +4
  • Research Article
  • Citations50

Unsupervised Person Re-Identification With Stochastic Training Strategy.

  • Jan 01, 2022
  • IEEE Transactions on Image Processing
  • Tianyang Liu +2
  • Research Article
  • Citations50

ST3D++: Denoised Self-Training for Unsupervised Domain Adaptation on 3D Object Detection.

  • Jan 01, 2022
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Jihan Yang +4
  • Conference Article
  • Citations41

Transformer Based Multi-Grained Features for Unsupervised Person Re-Identification

  • Jan 01, 2023
  • Jiachen Li +2
  • Conference Article
  • Citations26

Unsupervised Vehicle Re-Identification via Self-supervised Metric Learning using Feature Dictionary

  • Sep 27, 2021
  • Jongmin Yu +1
  • Research Article
  • Citations1

CLC²SIE: Cross-Level Consistency Constraint for Semisupervised Building Instance Extraction From High-Resolution Remote Sensing Imagery

  • Jan 01, 2025
  • IEEE Transactions on Geoscience and Remote Sensing
  • Rui Xu +5
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.