Identity-Guided Human Semantic Parsing for Person Re-Identification

TL;DR

This paper introduces an identity-guided semantic parsing method for person re-identification that locates body parts and belongings at pixel-level using only identity labels, improving alignment and recognition accuracy.

Contribution

The proposed ISP method uniquely generates pseudo-labels for human parts and belongings without pretrained models, enhancing person re-ID performance.

Findings

Outperforms state-of-the-art methods on three datasets

Effectively locates personal belongings like backpacks and reticules

Uses only identity labels for training, avoiding complex annotations

Abstract

Existing alignment-based methods have to employ the pretrained human parsing models to achieve the pixel-level alignment, and cannot identify the personal belongings (e.g., backpacks and reticule) which are crucial to person re-ID. In this paper, we propose the identity-guided human semantic parsing approach (ISP) to locate both the human body parts and personal belongings at pixel-level for aligned person re-ID only with person identity labels. We design the cascaded clustering on feature maps to generate the pseudo-labels of human parts. Specifically, for the pixels of all images of a person, we first group them to foreground or background and then group the foreground pixels to human parts. The cluster assignments are subsequently used as pseudo-labels of human parts to supervise the part estimation and ISP iteratively learns the feature maps and groups them. Finally, local features of both human body parts and personal belongings are obtained according to the selflearned part estimation, and only features of visible parts are utilized for the retrieval. Extensive experiments on three widely used datasets validate the superiority of ISP over lots of state-of-the-art methods. Our code is available at https://github.com/CASIA-IVA-Lab/ISP-reID.