Mining False Positive Examples for Text-Based Person Re-identification
This work addresses a challenging problem in cross-modal retrieval for security or surveillance applications, but it appears incremental as it builds on existing methods by focusing on mismatched pairs.
The paper tackles the problem of text-based person re-identification by addressing the neglect of mismatched word-region pairs, proposing a method that mines false positive examples and achieves superior effectiveness on the CUHK-PEDES dataset.
Text-based person re-identification (ReID) aims to identify images of the targeted person from a large-scale person image database according to a given textual description. However, due to significant inter-modal gaps, text-based person ReID remains a challenging problem. Most existing methods generally rely heavily on the similarity contributed by matched word-region pairs, while neglecting mismatched word-region pairs which may play a decisive role. Accordingly, we propose to mine false positive examples (MFPE) via a jointly optimized multi-branch architecture to handle this problem. MFPE contains three branches including a false positive mining (FPM) branch to highlight the role of mismatched word-region pairs. Besides, MFPE delicately designs a cross-relu loss to increase the gap of similarity scores between matched and mismatched word-region pairs. Extensive experiments on CUHK-PEDES demonstrate the superior effectiveness of MFPE. Our code is released at https://github.com/xx-adeline/MFPE.