A Re-ranking Method using K-nearest Weighted Fusion for Person Re-identification

2025-09-04

AI Summary

To address view bias in person re-identification caused by viewpoint, pose, and occlusion variations, this paper proposes an unsupervised K-Nearest Neighbor Weighted Fusion reranking method (KWF). Without model fine-tuning or additional annotations, KWF selects K nearest neighbors from initial retrieval results and performs weighted aggregation of their features via a learnable, similarity-driven weighting strategy to reconstruct robust multi-view representations. This effectively mitigates the viewpoint dependency inherent in single-view features, enhancing both ranking accuracy and generalization. On Market1501, MSMT17, and Occluded-DukeMTMC, KWF achieves Rank@1 scores of 96.3%, 72.1%, and 74.8%, respectively—improving Rank@1 by 9.8% and 22.0% over the baseline on MSMT17 and Occluded-DukeMTMC, with corresponding substantial gains in mAP. Moreover, KWF attains superior computational efficiency compared to existing reranking approaches.

Technology Category

• Computer Vision: Image and Video Retrieval
• Machine Learning: Multi-instance/Multi-view Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

In person re-identification, re-ranking is a crucial step to enhance the overall accuracy by refining the initial ranking of retrieved results. Previous studies have mainly focused on features from single-view images, which can cause view bias and issues like pose variation, viewpoint changes, and occlusions. Using multi-view features to present a person can help reduce view bias. In this work, we present an efficient re-ranking method that generates multi-view features by aggregating neighbors' features using K-nearest Weighted Fusion (KWF) method. Specifically, we hypothesize that features extracted from re-identification models are highly similar when representing the same identity. Thus, we select K neighboring features in an unsupervised manner to generate multi-view features. Additionally, this study explores the weight selection strategies during feature aggregation, allowing us to identify an effective strategy. Our re-ranking approach does not require model fine-tuning or extra annotations, making it applicable to large-scale datasets. We evaluate our method on the person re-identification datasets Market1501, MSMT17, and Occluded-DukeMTMC. The results show that our method significantly improves Rank@1 and mAP when re-ranking the top M candidates from the initial ranking results. Specifically, compared to the initial results, our re-ranking method achieves improvements of 9.8%/22.0% in Rank@1 on the challenging datasets: MSMT17 and Occluded-DukeMTMC, respectively. Furthermore, our approach demonstrates substantial enhancements in computational efficiency compared to other re-ranking methods.

Problem

• Reduces view bias in person re-identification using multi-view features
• Generates multi-view features via K-nearest weighted fusion method
• Improves re-ranking accuracy without model fine-tuning or extra annotations

Innovation

• K-nearest Weighted Fusion for feature aggregation
• Unsupervised multi-view feature generation
• Weight selection strategy exploration for optimization