# Who Remains, What Changes: Identity Anchored Composed Gait Retrieval

arXiv: https://arxiv.org/abs/2608.26632 · HTML: https://mlsanity.com/paper/2608.26632/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 39 references that are in the map.
- First posted on arXiv 27 Aug 2026.

## Builds on (most cited first, 21 of 21)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [Composing Text and Image for Image Retrieval - An Empirical Odyssey](https://mlsanity.com/paper/1812.07119/) (2018)
- [Zero-Shot Composed Image Retrieval with Textual Inversion](https://mlsanity.com/paper/2303.15247/) (2023)
- [Gait Recognition in the Wild with Dense 3D Representations and A Benchmark](https://mlsanity.com/paper/2204.02569/) (2022)
- [GaitSet: Cross-view Gait Recognition through Utilizing Gait as a Deep Set](https://mlsanity.com/paper/2102.03247/) (2021)
- [LidarGait: Benchmarking 3D Gait Recognition with Point Clouds](https://mlsanity.com/paper/2211.10598/) (2022)
- [Composed Image Retrieval using Contrastive Learning and Task-oriented CLIP-based Features](https://mlsanity.com/paper/2308.11485/) (2023)
- [CoVR-2: Automatic Data Construction for Composed Video Retrieval](https://mlsanity.com/paper/2308.14746/) (2023)
- [Exploring Deep Models for Practical Gait Recognition](https://mlsanity.com/paper/2303.03301/) (2023)
- [A Comprehensive Survey on Deep Gait Recognition: Algorithms, Datasets and Challenges](https://mlsanity.com/paper/2206.13732/) (2022)
- [BigGait: Learning Gait Representation You Want by Large Vision Models](https://mlsanity.com/paper/2402.19122/) (2024)
- [Parsing is All You Need for Accurate Gait Recognition in the Wild](https://mlsanity.com/paper/2308.16739/) (2023)
- [Cross-Covariate Gait Recognition: A Benchmark](https://mlsanity.com/paper/2312.14404/) (2023)
- [BiggerGait: Unlocking Gait Recognition with Layer-wise Representations from Large Vision Models](https://mlsanity.com/paper/2505.18132/) (2025)
- [Instance-Level Composed Image Retrieval](https://mlsanity.com/paper/2510.25387/) (2025)
- [GaitRef: Gait Recognition with Refined Sequential Skeletons](https://mlsanity.com/paper/2304.07916/) (2023)
- [HyCIR: Boosting Zero-Shot Composed Image Retrieval with Synthetic Labels](https://mlsanity.com/paper/2407.05795/) (2024)
- [Automatic Synthetic Data and Fine-grained Adaptive Feature Alignment for Composed Person Retrieval](https://mlsanity.com/paper/2311.16515/) (2023)
- [GaitSnippet: Gait Recognition Beyond Unordered Sets and Ordered Sequences](https://mlsanity.com/paper/2508.07782/) (2025)
- [Composed Image Retrieval for Training-Free Domain Conversion](https://mlsanity.com/paper/2412.03297/) (2024)

## Abstract (arXiv)

Gait recognition has achieved remarkable progress, yet existing methods remain confined to rigid visual matching and often overlook the potential of natural language instructions for interactive retrieval. In this paper, we introduce Composed Gait Retrieval (CoGR), a novel task that retrieves a target gait sequence based on a reference sequence and a natural language modification query. To address the absence of existing datasets for this task, we design an automated annotation pipeline powered by large vision-language models (VLMs) to construct the first gait-language datasets: Language-Augmented CCPG and Language-Augmented CASIA-B. Building on this, we propose ComposeGait, an identity-anchored composition framework designed to prevent the identity drift that arises when generic composed retrieval follows the instruction but returns the wrong person. Its Part-aware Identity Adapter (PIA) aggregates multi-frame, part-aware identity evidence into a sample-specific ID token. We inject the ID tokens into both branches of a shared Q-Former to preserve identity, while excluding the ID-token outputs from the final retrieval embeddings. Joint identity and task-adapted composed-retrieval objectives optimize this space end to end. We evaluate ComposeGait on both benchmarks and show that it achieves the best R@1 among the compared methods, reaching 72.38% on Language-Augmented CCPG and 83.61% on Language-Augmented CASIA-B. These results establish ComposeGait as a strong baseline for CoGR. The datasets and code will be made publicly available.