Papers › Prototypical Contrastive Learning-based CLIP Fine-tuning for Object Re-identification

Prototypical Contrastive Learning-based CLIP Fine-tuning for Object Re-identification

26 Oct 2023arXiv:2310.17218archive 2025-07-28

Jiachen Li, Xiaojin Gong

This work aims to adapt large-scale pre-trained vision-language models, such as contrastive language-image pretraining (CLIP), to enhance the performance of object reidentification (Re-ID) across various supervision settings. Although prompt learning has enabled a recent work named CLIP-ReID to achieve promising performance, the underlying mechanisms and the necessity of prompt learning remain unclear due to the absence of semantic labels in ReID tasks. In this work, we first analyze the role prompt learning in CLIP-ReID and identify its limitations. Based on our investigations, we propose a simple yet effective approach to adapt CLIP for supervised object Re-ID. Our approach directly fine-tunes the image encoder of CLIP using a prototypical contrastive learning (PCL) loss, eliminating the need for prompt learning. Experimental results on both person and vehicle Re-ID datasets demonstrate the competitiveness of our method compared to CLIP-ReID. Furthermore, we extend our PCL-based CLIP fine-tuning approach to unsupervised scenarios, where we achieve state-of-the art performance.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

RikoLi/PCL-CLIP officialmentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Contrastive LearningPerson Re-IdentificationPrompt LearningUnsupervised Person Re-IdentificationUnsupervised Vehicle Re-Identification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Person Re-Identification MSMT17 PCL-CLIP (L_pcl+L_id) Rank-1 89.8 #8 of 43 Archive leaderboard report
Person Re-Identification MSMT17 PCL-CLIP (L_pcl+L_id) Rank-10 96.0 #8 of 43 Archive leaderboard report
Person Re-Identification MSMT17 PCL-CLIP (L_pcl+L_id) Rank-5 94.7 #8 of 43 Archive leaderboard report
Person Re-Identification MSMT17 PCL-CLIP (L_pcl+L_id) mAP 76.1 #8 of 43 Archive leaderboard report
Person Re-Identification MSMT17 PCL-CLIP (L_pcl) Rank-1 89.2 #12 of 43 Archive leaderboard report
Person Re-Identification MSMT17 PCL-CLIP (L_pcl) Rank-10 95.8 #12 of 43 Archive leaderboard report
Person Re-Identification MSMT17 PCL-CLIP (L_pcl) Rank-5 94.7 #12 of 43 Archive leaderboard report
Person Re-Identification MSMT17 PCL-CLIP (L_pcl) mAP 73.8 #12 of 43 Archive leaderboard report
Person Re-Identification Market-1501 PCL-CLIP (L_pcl) Rank-1 96.1 #29 of 135 Archive leaderboard report
Person Re-Identification Market-1501 PCL-CLIP (L_pcl) Rank-5 98.8 #29 of 135 Archive leaderboard report
Person Re-Identification Market-1501 PCL-CLIP (L_pcl) mAP 91.0 #29 of 135 Archive leaderboard report
Person Re-Identification Market-1501 PCL-CLIP (L_pcl+L_id) Rank-1 95.9 #36 of 135 Archive leaderboard report
Person Re-Identification Market-1501 PCL-CLIP (L_pcl+L_id) Rank-5 98.5 #36 of 135 Archive leaderboard report
Person Re-Identification Market-1501 PCL-CLIP (L_pcl+L_id) mAP 91.4 #36 of 135 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (O2CAP) Rank-1 84.9 #1 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (O2CAP) Rank-10 94.0 #1 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (O2CAP) Rank-5 92.0 #1 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (O2CAP) mAP 65.5 #1 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CAP) Rank-1 79.0 #3 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CAP) Rank-10 91.1 #3 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CAP) Rank-5 88.4 #3 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CAP) mAP 53.6 #3 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CC) Rank-1 77.9 #4 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CC) Rank-10 87.2 #4 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CC) Rank-5 85.2 #4 of 12 Archive leaderboard report
Unsupervised Person Re-Identification MSMT17 PCL-CLIP (CC) mAP 56.4 #4 of 12 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (O2CAP) MAP 88.4 #3 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (O2CAP) Rank-1 94.8 #3 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (O2CAP) Rank-10 98.7 #3 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (O2CAP) Rank-5 98.0 #3 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CAP) MAP 87.4 #5 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CAP) Rank-1 93.9 #5 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CAP) Rank-10 98.5 #5 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CAP) Rank-5 97.7 #5 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CC) MAP 86.9 #6 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CC) Rank-1 94.2 #6 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CC) Rank-10 98.7 #6 of 23 Archive leaderboard report
Unsupervised Person Re-Identification Market-1501 PCL-CLIP (CC) Rank-5 97.8 #6 of 23 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

CLIPContrastive Learning

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections