Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification

2025-01-01

Italian National Conference on Sensors

AI Summary

Unsupervised domain-generalizable person re-identification (DG Re-ID) suffers from models over-relying on domain-biased shortcuts, leading to poor generalization. Method: We propose Diffusion-assisted Correlation-Aware Representation Learning (DCAC), the first framework to integrate pre-trained diffusion models (e.g., Stable Diffusion) into DG Re-ID. DCAC introduces learnable ID-prompt embeddings guided by ID classification probabilities and a correlation-aware conditional mechanism, enabling bidirectional co-optimization between discriminative/contrastive Re-ID features and the diffusion generative process. Through probability-guided conditional control and gradient-based backward feedback, DCAC explicitly injects identity-relevant latent knowledge to mitigate shortcut learning. Contribution/Results: DCAC achieves state-of-the-art performance on both single-source and multi-source unsupervised DG Re-ID benchmarks. Ablation studies confirm the effectiveness of each component in enhancing cross-domain generalization.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Learning to Search
• Machine Learning: Deep Generative Models & Autoencoders

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: Graph embeddings and representation learning for Web-related graphs

Abstract

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.

Problem

• Enhance domain-generalizable person re-identification
• Mitigate shortcut learning in Re-ID models
• Integrate diffusion models with Re-ID frameworks

Innovation

• pre-trained diffusion models
• correlation-aware conditioning scheme
• enhanced generalizable feature representation