Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification
Domain-generalizable re-identification (DG Re-ID) aims to train a model on one or more source domains and evaluate its performance on unseen target domains, a task that has attracted growing attention due to its practical relevance. While numerous methods have been proposed, most rely on discriminat...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/25/2/552 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832587514227982336 |
|---|---|
| author | Jiachen Li Xiaojin Gong |
| author_facet | Jiachen Li Xiaojin Gong |
| author_sort | Jiachen Li |
| collection | DOAJ |
| description | Domain-generalizable re-identification (DG Re-ID) aims to train a model on one or more source domains and evaluate its performance on unseen target domains, a task that has attracted growing attention due to its practical relevance. While numerous methods have been proposed, most rely on discriminative or contrastive learning frameworks to learn generalizable feature representations. However, these approaches often fail to mitigate shortcut learning, leading to suboptimal performance. In this work, we propose a novel method called diffusion model-assisted representation learning with a correlation-aware conditioning scheme (DCAC) to enhance DG Re-ID. Our method integrates a discriminative and contrastive Re-ID model with a pre-trained diffusion model through a correlation-aware conditioning scheme. By incorporating ID classification probabilities generated from the Re-ID model with a set of learnable ID-wise prompts, the conditioning scheme injects dark knowledge that captures ID correlations to guide the diffusion process. Simultaneously, feedback from the diffusion model is back-propagated through the conditioning scheme to the Re-ID model, effectively improving the generalization capability of Re-ID features. Extensive experiments on both single-source and multi-source DG Re-ID tasks demonstrate that our method achieves state-of-the-art performance. Comprehensive ablation studies further validate the effectiveness of the proposed approach, providing insights into its robustness. |
| format | Article |
| id | doaj-art-9c8e695ada9f4e5888b9868e6b151dee |
| institution | Kabale University |
| issn | 1424-8220 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj-art-9c8e695ada9f4e5888b9868e6b151dee2025-01-24T13:49:19ZengMDPI AGSensors1424-82202025-01-0125255210.3390/s25020552Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-IdentificationJiachen Li0Xiaojin Gong1College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, ChinaCollege of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, ChinaDomain-generalizable re-identification (DG Re-ID) aims to train a model on one or more source domains and evaluate its performance on unseen target domains, a task that has attracted growing attention due to its practical relevance. While numerous methods have been proposed, most rely on discriminative or contrastive learning frameworks to learn generalizable feature representations. However, these approaches often fail to mitigate shortcut learning, leading to suboptimal performance. In this work, we propose a novel method called diffusion model-assisted representation learning with a correlation-aware conditioning scheme (DCAC) to enhance DG Re-ID. Our method integrates a discriminative and contrastive Re-ID model with a pre-trained diffusion model through a correlation-aware conditioning scheme. By incorporating ID classification probabilities generated from the Re-ID model with a set of learnable ID-wise prompts, the conditioning scheme injects dark knowledge that captures ID correlations to guide the diffusion process. Simultaneously, feedback from the diffusion model is back-propagated through the conditioning scheme to the Re-ID model, effectively improving the generalization capability of Re-ID features. Extensive experiments on both single-source and multi-source DG Re-ID tasks demonstrate that our method achieves state-of-the-art performance. Comprehensive ablation studies further validate the effectiveness of the proposed approach, providing insights into its robustness.https://www.mdpi.com/1424-8220/25/2/552generalizable person re-identificationdiffusion modelprompt learningCLIP |
| spellingShingle | Jiachen Li Xiaojin Gong Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification Sensors generalizable person re-identification diffusion model prompt learning CLIP |
| title | Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification |
| title_full | Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification |
| title_fullStr | Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification |
| title_full_unstemmed | Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification |
| title_short | Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification |
| title_sort | unleashing the potential of pre trained diffusion models for generalizable person re identification |
| topic | generalizable person re-identification diffusion model prompt learning CLIP |
| url | https://www.mdpi.com/1424-8220/25/2/552 |
| work_keys_str_mv | AT jiachenli unleashingthepotentialofpretraineddiffusionmodelsforgeneralizablepersonreidentification AT xiaojingong unleashingthepotentialofpretraineddiffusionmodelsforgeneralizablepersonreidentification |