# From Global to Local: Rethinking CLIP Feature Aggregation for Person Re-Identification

arXiv: https://arxiv.org/abs/2604.22190 · HTML: https://mlsanity.com/paper/2604.22190/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 22 references that are in the map.
- First posted on arXiv 24 Apr 2026.

## Builds on (most cited first, 19 of 19)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Moment Matching for Multi-Source Domain Adaptation](https://mlsanity.com/paper/1812.01754/) (2018)
- [Perceiver: General Perception with Iterative Attention](https://mlsanity.com/paper/2103.03206/) (2021)
- [Object-Centric Learning with Slot Attention](https://mlsanity.com/paper/2006.15055/) (2020)
- [In Defense of the Triplet Loss for Person Re-Identification](https://mlsanity.com/paper/1703.07737/) (2017)
- [Deep Learning for Person Re-identification: A Survey and Outlook](https://mlsanity.com/paper/2001.04193/) (2020)
- [Beyond Part Models: Person Retrieval with Refined Part Pooling (and a Strong Convolutional Baseline)](https://mlsanity.com/paper/1711.09349/) (2017)
- [TransReID: Transformer-based Object Re-Identification](https://mlsanity.com/paper/2102.04378/) (2021)
- [Bag of Tricks and A Strong Baseline for Deep Person Re-identification](https://mlsanity.com/paper/1903.07071/) (2019)
- [Person Re-identification: Past, Present and Future](https://mlsanity.com/paper/1610.02984/) (2016)
- [CLIP-ReID: Exploiting Vision-Language Model for Image Re-Identification without Concrete Text Labels](https://mlsanity.com/paper/2211.13977/) (2022)
- [Identity-Guided Human Semantic Parsing for Person Re-Identification](https://mlsanity.com/paper/2007.13467/) (2020)
- [ERM++: An Improved Baseline for Domain Generalization](https://mlsanity.com/paper/2304.01973/) (2023)
- [Feature Completion Transformer for Occluded Person Re-identification](https://mlsanity.com/paper/2303.01656/) (2023)
- [ProFD: Prompt-Guided Feature Disentangling for Occluded Person Re-Identification](https://mlsanity.com/paper/2409.20081/) (2024)
- [Multi-modal Multi-platform Person Re-Identification: Benchmark and Method](https://mlsanity.com/paper/2503.17096/) (2025)
- [Parallel Augmentation and Dual Enhancement for Occluded Person Re-identification](https://mlsanity.com/paper/2210.05438/) (2022)
- [CLIP-SCGI: Synthesized Caption-Guided Inversion for Person Re-Identification](https://mlsanity.com/paper/2410.09382/) (2024)
- [SCING:Towards More Efficient and Robust Person Re-Identification through Selective Cross-modal Prompt Tuning](https://mlsanity.com/paper/2507.00506/) (2025)

## Abstract (arXiv)

CLIP-based person re-identification (ReID) methods aggregate spatial features into a single global texttt[CLS] token optimized for image-text alignment rather than spatial selectivity, making representations fragile under occlusion and cross-camera variation. We propose SAGA-ReID, which reconstructs identity representations by aligning intermediate patch tokens with anchor vectors parameterized in CLIP's text embedding space -- emphasizing spatially stable evidence while suppressing corrupted or absent regions, without requiring textual descriptions of individual images. Controlled experiments isolate the aggregation mechanism under two qualitatively distinct conditions -- synthetic masking, where identity signal is absent, and realistic human distractors, where an overlapping person introduces semantically confusing signal -- with SAGA's advantage over global pooling growing substantially as occlusion increases across both conditions. Benchmark evaluations confirm consistent gains over CLIP-ReID across standard and occluded settings, with the largest improvements where global pooling is most unreliable: up to +10.6 Rank-1 on occluded benchmarks. SAGA's aggregation outperforms dedicated sequential patch aggregation on a stronger backbone, confirming that structured reconstruction addresses a bottleneck that backbone quality and architectural complexity alone cannot resolve. Code available at https://github.com/ipl-uw/Structured-Anchor-Guided-Aggregation-for-ReID.