# ITSELF: Attention Guided Fine-Grained Alignment for Vision-Language Retrieval

arXiv: https://arxiv.org/abs/2601.01024 · HTML: https://mlsanity.com/paper/2601.01024/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 35 of its 48 references that are in the map.
- First posted on arXiv 3 Jan 2026.

## Placed in

- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)

## Builds on (most cited first, 30 of 35)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Align before Fuse: Vision and Language Representation Learning with Momentum Distillation](https://mlsanity.com/paper/2107.07651/) (2021)
- [Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned](https://mlsanity.com/paper/1905.09418/) (2019)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [Person Search with Natural Language Description](https://mlsanity.com/paper/1702.05729/) (2017)
- [Learning the Best Pooling Strategy for Visual Semantic Embedding](https://mlsanity.com/paper/2011.04305/) (2020)
- [Cross-Modal Implicit Relation Reasoning and Aligning for Text-to-Image Person Retrieval](https://mlsanity.com/paper/2303.12501/) (2023)
- [Semantically Self-Aligned Network for Text-to-Image Part-aware Person Re-identification](https://mlsanity.com/paper/2107.12666/) (2021)
- [DSSL: Deep Surroundings-person Separation Learning for Text-based Person Retrieval](https://mlsanity.com/paper/2109.05534/) (2021)
- [CLIP-Driven Fine-grained Text-Image Person Re-identification](https://mlsanity.com/paper/2210.10276/) (2022)
- [ViTAA: Visual-Textual Attributes Alignment in Person Search by Natural Language](https://mlsanity.com/paper/2005.07327/) (2020)
- [Towards Unified Text-based Person Retrieval: A Large-scale Multi-Attribute and Language Search Benchmark](https://mlsanity.com/paper/2306.02898/) (2023)
- [See Finer, See More: Implicit Modality Alignment for Text-based Person Retrieval](https://mlsanity.com/paper/2208.08608/) (2022)
- [Learning Granularity-Unified Representations for Text-to-Image Person Re-identification](https://mlsanity.com/paper/2207.07802/) (2022)
- [Noisy-Correspondence Learning for Text-to-Image Person Re-identification](https://mlsanity.com/paper/2308.09911/) (2023)
- [RaSa: Relation and Sensitivity Aware Representation Learning for Text-based Person Search](https://mlsanity.com/paper/2305.13653/) (2023)
- [PLIP: Language-Image Pre-training for Person Representation Learning](https://mlsanity.com/paper/2305.08386/) (2023)
- [Harnessing the Power of MLLMs for Transferable Text-to-Image Person ReID](https://mlsanity.com/paper/2405.04940/) (2024)
- [Image-Specific Information Suppression and Implicit Local Alignment for Text-based Person Search](https://mlsanity.com/paper/2208.14365/) (2022)
- [An Empirical Study of CLIP for Text-based Person Search](https://mlsanity.com/paper/2308.10045/) (2023)
- [AXM-Net: Implicit Cross-Modal Feature Alignment for Person Re-identification](https://mlsanity.com/paper/2101.08238/) (2021)
- [UFineBench: Towards Text-based Person Retrieval with Ultra-fine Granularity](https://mlsanity.com/paper/2312.03441/) (2023)
- [UniPT: Universal Parallel Tuning for Transfer Learning with Efficient Parameter and Memory](https://mlsanity.com/paper/2308.14316/) (2023)
- [Unified Pre-training with Pseudo Texts for Text-To-Image Person Re-identification](https://mlsanity.com/paper/2309.01420/) (2023)
- [PLOT: Text-based Person Search with Part Slot Attention for Corresponding Part Discovery](https://mlsanity.com/paper/2409.13475/) (2024)
- [Improving Generalization in Visual Reasoning via Self-Ensemble](https://mlsanity.com/paper/2410.20883/) (2024)
- [MARS: Paying more attention to visual attributes for text-based person search](https://mlsanity.com/paper/2407.04287/) (2024)
- [CLIP-based Synergistic Knowledge Transfer for Text-based Person Retrieval](https://mlsanity.com/paper/2309.09496/) (2023)
- [DM-Adapter: Domain-Aware Mixture-of-Adapters for Text-Based Person Retrieval](https://mlsanity.com/paper/2503.04144/) (2025)

## Built on by (most cited first, 3 of 3)

- [Harnessing Weak Pair Uncertainty for Text-based Person Search](https://mlsanity.com/paper/2604.08877/) (2026)
- [ESC: Emotional Self-Correction for Reliable Vision-Language Models](https://mlsanity.com/paper/2607.02089/) (2026)
- [Heterogeneous Vision-Language Ensemble with Disagreement-Aware Reranking for Text-Based Person Anomaly Retrieval](https://mlsanity.com/paper/2608.12843/) (2026)

## Abstract (arXiv)

Vision Language Models (VLMs) have rapidly advanced and show strong promise for text-based person search (TBPS), a task that requires capturing fine-grained relationships between images and text to distinguish individuals. Previous methods address these challenges through local alignment, yet they are often prone to shortcut learning and spurious correlations, yielding misalignment. Moreover, injecting prior knowledge can distort intra-modality structure. Motivated by our finding that encoder attention surfaces spatially precise evidence from the earliest training epochs, and to alleviate these issues, we introduceITSELF, an attention-guided framework for implicit local alignment. At its core, Guided Representation with Attentive Bank (GRAB) converts the model's own attention into an Attentive Bank of high-saliency tokens and applies local objectives on this bank, learning fine-grained correspondences without extra supervision. To make the selection reliable and non-redundant, we introduce Multi-Layer Attention for Robust Selection (MARS), which aggregates attention across layers and performs diversity-aware top-k selection; and Adaptive Token Scheduler (ATS), which schedules the retention budget from coarse to fine over training, preserving context early while progressively focusing on discriminative details. Extensive experiments on three widely used TBPS benchmarks showstate-of-the-art performance and strong cross-dataset generalization, confirming the effectiveness and robustness of our approach without additional prior supervision. Our project is publicly available at https://trhuuloc.github.io/itself