KPDM: Key Phrase Dynamic Masking for Robust Text-to-Image Person Retrieval
Shaofeng You, Tianle Miao, Qihang Chen, Xin Li, Zhuo Cheng, Dapeng Luo
Abstract
Text-to-image person re-identification (TIReID) aims to retrieve the most relevant pedestrian images from an image gallery based on natural language descriptions. Recent studies have achieved significant performance improvements by leveraging Masked Language Modeling (MLM) to align fine-grained information through local matching. However, in the text feature extraction, randomly masking text tokens may disrupt the semantic relationships between these local tokens, leading to feature misalignment; on the other hand, from an image feature perspective, redundant patches in pedestrian images hinder the information interaction across modalities. Moreover, the presence of noisy image-text pairs further complicates the learning process, as the model may be misled into recognizing incorrect patterns. To address these issues, we propose a robust fine-grained local alignment framework based on Key Phrase Dynamic Mask (KPDM). First, we strengthen the semantic relationships between text tokens by implementing a "adjective + noun
BibTeX
@inproceedings{aaai2026_kpdmkeyphrasedyn,
title = {KPDM: Key Phrase Dynamic Masking for Robust Text-to-Image Person Retrieval},
author = {Shaofeng You and Tianle Miao and Qihang Chen and Xin Li and Zhuo Cheng and Dapeng Luo},
booktitle = {AAAI 2026},
year = {2026}
}