Subtask-dominated Transfer Learning for Long-tail Person Search

TL;DR

This paper introduces a Subtask-dominated Transfer Learning approach to address long-tail identity distribution challenges in one-step person search, enhancing feature discrimination and overall performance.

Contribution

It proposes a novel transfer learning method focusing on the Re-ID subtask and a Multi-level RoI Fusion Pooling layer for improved person search accuracy.

Findings

Outperforms existing methods on CUHK-SYSU and PRW datasets.

Effectively mitigates long-tail identity distribution issues.

Enhances person feature discrimination in one-step search.

Abstract

Person search unifies person detection and person re-identification (Re-ID) to locate query persons from the panoramic gallery images. One major challenge comes from the imbalanced long-tail person identity distributions, which prevents the one-step person search model from learning discriminative person features for the final re-identification. However, it is under-explored how to solve the heavy imbalanced identity distributions for the one-step person search. Techniques designed for the long-tail classification task, for example, image-level re-sampling strategies, are hard to be effectively applied to the one-step person search which jointly solves person detection and Re-ID subtasks with a detection-based multi-task framework. To tackle this problem, we propose a Subtask-dominated Transfer Learning (STL) method. The STL method solves the long-tail problem in the pretraining stage of the dominated Re-ID subtask and improves the one-step person search by transfer learning of the pretrained model. We further design a Multi-level RoI Fusion Pooling layer to enhance the discrimination ability of person features for the one-step person search. Extensive experiments on CUHK-SYSU and PRW datasets demonstrate the superiority and effectiveness of the proposed method.