Imitating Targets from all sides: An Unsupervised Transfer Learning method for Person Re-identification

TL;DR

This paper introduces an unsupervised transfer learning approach for person re-identification that reduces inter-dataset bias and intra-dataset differences by leveraging a dual classification loss and class-style space commonality, improving cross-domain generalization.

Contribution

It proposes a novel ImitateModel and a dual classification loss to effectively transfer knowledge across datasets in an unsupervised manner for person Re-ID.

Findings

Achieves competitive performance on Market-1501 and DukeMTMC-reID benchmarks.

Effectively reduces inter-dataset bias and intra-dataset differences.

Demonstrates improved generalization in unsupervised person Re-ID.

Abstract

Person re-identification (Re-ID) models usually show a limited performance when they are trained on one dataset and tested on another dataset due to the inter-dataset bias (e.g. completely different identities and backgrounds) and the intra-dataset difference (e.g. camera invariance). In terms of this issue, given a labelled source training set and an unlabelled target training set, we propose an unsupervised transfer learning method characterized by 1) bridging inter-dataset bias and intra-dataset difference via a proposed ImitateModel simultaneously; 2) regarding the unsupervised person Re-ID problem as a semi-supervised learning problem formulated by a dual classification loss to learn a discriminative representation across domains; 3) exploiting the underlying commonality across different domains from the class-style space to improve the generalization ability of re-ID models. Extensive experiments are conducted on two widely employed benchmarks, including Market-1501 and DukeMTMC-reID, and experimental results demonstrate that the proposed method can achieve a competitive performance against other state-of-the-art unsupervised Re-ID approaches.