通过几何约束建模自适应衣物子空间，提升换衣场景下的人体重识别准确率。
Learning Instance-Adaptive Low-Rank Orthogonal Subspaces for Clothes-Changing Person Re-Identification
- 用文本描述生成低秩衣物子空间，通过交叉注意力实现个体自适应
- 在PRCC等数据集上提升5.9%~5.3%，显著优于现有方法
- 适合处理衣物变化大、可见度不一的现实重识别场景
换衣人体重识别旨在应对服装变化导致的外观剧烈变动。现有方法依赖对抗学习分离衣物特征，本文提出Ortho-ReID，通过视觉语言模型文本描述显式构建低秩衣物子空间，并利用直接几何约束提取衣物不变表征。核心组件为基于Transformer的Basis Maker，通过图像块与文本的交叉注意力，将共享的低维衣物先验细化为个体自适应的低秩子空间，即使在可视性变化条件下仍能稳健提取衣物特征。该子空间通过与衣物文本嵌入对齐进行监督，身份特征则通过可学习投影头提取，并被几何约束为严格正交于衣物子空间。大量实验表明，在PRCC（+5.9% top-1）、Celeb-reID-light（+3.5%）和LaST（+5.3%）上达到最先进性能，且在LTCC上表现优异。
原文摘要 · Abstract (English)
Clothes-changing person re-identification (CC-ReID) aims to recognize individuals despite drastic appearance changes caused by clothing variation. While existing methods rely on adversarial learning to disentangle clothing features, we propose Ortho-ReID, which explicitly models a low-rank clothing subspace from VLM text descriptions and extracts clothing-invariant representations via direct geometric constraints. A critical component is our transformer-based Basis Maker, which refines a shared, low-dimensional clothing prior into an instance-adaptive low-rank subspace through cross-attention with image patches, enabling robust clothing feature extraction even under varying visibility conditions. This instance-adaptive subspace is supervised via alignment with clothing text embeddings, while identity features are extracted via a learnable projection head and geometrically constrained to be strictly orthogonal to it. Extensive experiments demonstrate state-of-the-art performance on PRCC (+5.9% top-1), Celeb-reID-light (+3.5%), and LaST (+5.3%), with competitive results on LTCC.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。