Generalizable Person Re-identification via Balancing Alignment and Uniformity

TL;DR

A novel framework, Balancing Alignment and Uniformity (BAU), is proposed, which effectively mitigates this effect by maintaining a balance between alignment and uniformity and introduces a domain-specific uniformity loss that promotes uniformity within each source domain, thereby enhancing the learning of domain-invariant features.

Abstract

Domain generalizable person re-identification (DG re-ID) aims to learn discriminative representations that are robust to distributional shifts. While data augmentation is a straightforward solution to improve generalization, certain augmentations exhibit a polarized effect in this task, enhancing in-distribution performance while deteriorating out-of-distribution performance. In this paper, we investigate this phenomenon and reveal that it leads to sparse representation spaces with reduced uniformity. To address this issue, we propose a novel framework, Balancing Alignment and Uniformity (BAU), which effectively mitigates this effect by maintaining a balance between alignment and uniformity. Specifically, BAU incorporates alignment and uniformity losses applied to both original and augmented images and integrates a weighting strategy to assess the reliability of augmented samples, further improving the alignment loss. Additionally, we introduce a domain-specific uniformity loss that promotes uniformity within each source domain, thereby enhancing the learning of domain-invariant features. Extensive experimental results demonstrate that BAU effectively exploits the advantages of data augmentation, which previous studies could not fully utilize, and achieves state-of-the-art performance without requiring complex training procedures. The code is available at \url{https://github.com/yoonkicho/BAU}.

Figures (8)

• Figure 1: Analysis on polarized effect of data augmentations on in-distribution (ID) and out-of-distribution (OOD). (a) mAP (%) on Market-1501 of models trained on the same dataset (ID) and MS+CS+C3 (OOD) with varying augmentation probabilities. (b) Alignment ($\mathcal{L}_{\text{align}}$) and uniformity ($\mathcal{L}_{\text{uniform}}$) of OOD scenarios (MS+CS+C3 $\rightarrow$ M). Counterintuitively, augmentations lead to more alignment but less uniformity, indicating that the model fails to sufficiently preserve the diverse information from the data distribution. (c) Uniformity ($- \mathcal{L}_{\text{uniform}}$) vs. augmentation probability for the source and target datasets in MS+CS+C3 $\rightarrow$ M. Higher probabilities result in less uniformity, especially under distribution shifts, indicating an insufficiency in representing OOD data.
• Figure 2: Grad-CAM selvaraju2017grad across different probabilities of data augmentations.
• Figure 3: Overview of the proposed framework. In (b) and (c), each color represents a different identity and domain, respectively. (a) With original and augmented images, we apply alignment and uniformity losses to balance feature discriminability and generalization capability. We further introduce a domain-specific uniformity loss to mitigate domain bias. (b) $\mathcal{L}_{\mathrm{align}}$ pulls positive features closer, while $\mathcal{L}_{\mathrm{uniform}}$ pushes all features apart to maintain diversity. (c) $\mathcal{L}_{\mathrm{domain}}$ uniformly distributes each domain's features and prototypes, reducing domain bias and thus enhancing generalization.
• Figure 4: Analysis of alignment and uniformity. (a) Alignment ($\mathcal{L}_{\mathrm{align}}$) and uniformity ($\mathcal{L}_{\mathrm{uniform}}$) on Market-1501 when MS+CS+C3 $\rightarrow$ M under Protocol-3 with varying augmentation probabilities. (b) T-SNE visualization with and without the domain-specific uniformity loss $\mathcal{L}_{\mathrm{domain}}$. The values in parentheses in each legend label indicate the uniformity of the corresponding domain.
• Figure 5: Analysis of the weighting strategy. (a) Quantitative comparison of mAP (%) across varying augmentation probabilities, with and without the weighting strategy, on MS+CS+C3 $\rightarrow$ M under Protocol-3. The weighting strategy consistently improves performance, especially at higher augmentation probabilities, where the mAP drops significantly without it. (b) Qualitative analysis of the weight score $w$ for different pairs of original and augmented images.