SCPNet: Spatial-Channel Parallelism Network for Joint Holistic and Partial Person Re-Identification

TL;DR

SCPNet introduces a spatial-channel parallelism approach that enhances both holistic and partial person re-identification by learning discriminative features through spatial-channel correspondence supervision, achieving competitive and state-of-the-art results.

Contribution

The paper proposes SCPNet, a novel network that jointly learns holistic and partial person re-identification using spatial-channel parallelism supervision.

Findings

Achieves competitive accuracy on four holistic ReID datasets.

Outperforms state-of-the-art on two partial ReID datasets without additional training.

Demonstrates effectiveness of spatial-channel supervision in ReID tasks.

Abstract

Holistic person re-identification (ReID) has received extensive study in the past few years and achieves impressive progress. However, persons are often occluded by obstacles or other persons in practical scenarios, which makes partial person re-identification non-trivial. In this paper, we propose a spatial-channel parallelism network (SCPNet) in which each channel in the ReID feature pays attention to a given spatial part of the body. The spatial-channel corresponding relationship supervises the network to learn discriminative feature for both holistic and partial person re-identification. The single model trained on four holistic ReID datasets achieves competitive accuracy on these four datasets, as well as outperforms the state-of-the-art methods on two partial ReID datasets without training.