AI RESEARCH
Towards Robust Text-to-Image Person Retrieval: Multi-View Reformulation for Semantic Compensation
arXiv CS.CV
•
ArXi:2604.18376v1 Announce Type: new In text-to-image person retrieval tasks, the diversity of natural language expressions and the implicitness of visual semantics often lead to the problem of Expression Drift, where semantically equivalent texts exhibit significant feature discrepancies in the embedding space due to phrasing variations, thereby degrading the robustness of image-text alignment.