Bridging the Source-to-target Gap for Cross-domain Person Re-Identification with Intermediate Domains

Abstract

Cross-domain person re-identification (re-ID), such as unsupervised domain adaptive re-ID (UDA re-ID), aims to transfer the identity-discriminative knowledge from the source to the target domain. Existing methods commonly consider the source and target domains are isolated from each other, i.e., no intermediate status is modeled between the source and target domains. Directly transferring the knowledge between two isolated domains can be very difficult, especially when the domain gap is large. This paper, from a novel perspective, assumes these two domains are not completely isolated, but can be connected through a series of intermediate domains. Instead of directly aligning the source and target domains against each other, we propose to align the source and target domains against their intermediate domains so as to facilitate a smooth knowledge transfer. To discover and utilize these intermediate domains, this paper proposes an Intermediate Domain Module (IDM) and a Mirrors Generation Module (MGM). IDM has two functions: 1) it generates multiple intermediate domains by mixing the hidden-layer features from source and target domains and 2) it dynamically reduces the domain gap between the source / target domain features and the intermediate domain features. While IDM achieves good domain alignment effect, it introduces a side effect, i.e., the mix-up operation may mix the identities into a new identity and lose the original identities. Accordingly, MGM is introduced to compensate the loss of the original identity by mapping the features into the IDM-generated intermediate domains without changing their original identity. It allows to focus on minimizing domain variations to further promote the alignment between the source / target domain and intermediate domains, which reinforces IDM into IDM++. We extensively evaluate our method under both the UDA and domain generalization (DG) scenarios and observe that IDM++ yields consistent (and usually significant) performance improvement for cross-domain re-ID, achieving new state of the art. For example, on the challenging MSMT17 benchmark, IDM++ surpasses the prior state of the art by a large margin (e.g., up to 9.9% and 7.8% rank-1 accuracy) for UDA and DG scenarios, respectively. Code will be publicly available.

Person re-identification accuracy improvement by training a CNN with the new large joint dataset and re-rank

“…Similarly, in [14], the use of target domain images for training allows increasing mAP for Market-1501 from 33.9% to 82.3% and for DukeMTMC-ReID from 33.6% to 73.2%. Maximum accuracy in mAP among the considered algorithms was achieved using the IDM algorithm [3,4] for the Market-1501 and MSMT17 datasets, which became possible using the features of intermediate domains generated during the learning process.…”

A Video Target Re-Recognition Method Based on Adaptive Attention Enhancement and Multi-Scale Feature Fusion

“…In this context, target re-recognition techniques for multiple video frames have gradually become a research hotspot in the field of computer vision. Target re-recognition aims to recognize and match the same target in different camera views, different times and environments, which is crucial for cross-camera target tracking, people retrieval and other applications [3][4][5]. Target re-recognition technology has a wide range of potential applications in public safety, smart city building, retail and many other fields.…”