Open-Attribute Recognition for Person Retrieval: Finding People Through Distinctive and Novel Attributes

2025-08-02

AI Summary

Existing pedestrian attribute recognition (PAR) methods rely on a closed-set assumption, limiting their ability to recognize unseen attributes at inference time and suffering from poor generalizability and discriminability of predefined attributes. To address this, we propose *Open-Attribute Recognition*—a new task enabling cross-scenario pedestrian retrieval grounded in both seen and unseen attribute cues. We introduce the first dedicated framework for this task: a decoupled body-part representation learning mechanism that jointly integrates attribute semantic embeddings with fine-grained regional modeling, thereby supporting generalization to novel and fine-grained attributes. Furthermore, we reconstruct four major benchmarks—including Market-1501—to establish a standardized open-attribute evaluation protocol. Extensive experiments demonstrate significant improvements in both open-world attribute recognition and downstream pedestrian retrieval performance. The code and models will be publicly released.

Technology Category

• Computer Vision: Image and Video Retrieval
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Knowledge Representation and Reasoning: Argumentation

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Pedestrian Attribute Recognition (PAR) plays a crucial role in various vision tasks such as person retrieval and identification. Most existing attribute-based retrieval methods operate under the closed-set assumption that all attribute classes are consistently available during both training and inference. However, this assumption limits their applicability in real-world scenarios where novel attributes may emerge. Moreover, predefined attributes in benchmark datasets are often generic and shared across individuals, making them less discriminative for retrieving the target person. To address these challenges, we propose the Open-Attribute Recognition for Person Retrieval (OAPR) task, which aims to retrieve individuals based on attribute cues, regardless of whether those attributes were seen during training. To support this task, we introduce a novel framework designed to learn generalizable body part representations that cover a broad range of attribute categories. Furthermore, we reconstruct four widely used datasets for open-attribute recognition. Comprehensive experiments on these datasets demonstrate the necessity of the OAPR task and the effectiveness of our framework. The source code and pre-trained models will be publicly available upon publication.

Problem

• Retrieving people using unseen attributes during training
• Overcoming generic predefined attributes in datasets
• Learning generalizable body part representations

Innovation

• Open-Attribute Recognition for novel attributes
• Generalizable body part representations learning
• Reconstructed datasets for open-attribute evaluation