Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification

TL;DR

This paper introduces a novel domain adaptation framework for person re-identification that preserves image self-similarity and domain-dissimilarity during translation, improving cross-domain re-ID accuracy.

Contribution

It proposes the SPGAN model combining Siamese and CycleGAN networks to maintain discriminative features and domain differences during image translation.

Findings

SPGAN-generated images improve domain adaptation performance.

The method achieves competitive re-ID accuracy on large datasets.

Preserving self-similarity and dissimilarity enhances translation quality.

Abstract

Person re-identification (re-ID) models trained on one domain often fail to generalize well to another. In our attempt, we present a "learning via translation" framework. In the baseline, we translate the labeled images from source to target domain in an unsupervised manner. We then train re-ID models with the translated images by supervised methods. Yet, being an essential part of this framework, unsupervised image-image translation suffers from the information loss of source-domain labels during translation. Our motivation is two-fold. First, for each image, the discriminative cues contained in its ID label should be maintained after translation. Second, given the fact that two domains have entirely different persons, a translated image should be dissimilar to any of the target IDs. To this end, we propose to preserve two types of unsupervised similarities, 1) self-similarity of an image before and after translation, and 2) domain-dissimilarity of a translated source image and a target image. Both constraints are implemented in the similarity preserving generative adversarial network (SPGAN) which consists of an Siamese network and a CycleGAN. Through domain adaptation experiment, we show that images generated by SPGAN are more suitable for domain adaptation and yield consistent and competitive re-ID accuracy on two large-scale datasets.