Deep Hashing with Semantic Hash Centers for Image Retrieval

2025-07-11

AI Summary

Existing deep hashing methods predefine data-agnostic hash centers, neglecting inter-class semantic relationships and thereby limiting retrieval performance. To address this, we propose a three-stage semantic-aware deep hashing framework. First, we introduce the concept of “semantic hash centers,” modeling inter-class semantic similarities via a classification network and enforcing a minimum-distance constraint to preserve semantic structure. Second, we optimize these centers in a data-driven manner. Third, we jointly train a deep hashing network to learn discriminative binary codes. This work is the first to explicitly embed semantic structure into hash center design. Extensive experiments on multiple benchmark datasets demonstrate significant improvements in retrieval accuracy: mean average precision (MAP) at 100, 1000, and all returned samples increases by 7.26%, 7.62%, and 11.71%, respectively, validating both the effectiveness and generalizability of semantic center modeling.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Computer Vision: Image and Video Retrieval
• Data Mining & Knowledge Management: Semantic Web

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Deep hashing is an effective approach for large-scale image retrieval. Current methods are typically classified by their supervision types: point-wise, pair-wise, and list-wise. Recent point-wise techniques (e.g., CSQ, MDS) have improved retrieval performance by pre-assigning a hash center to each class, enhancing the discriminability of hash codes across various datasets. However, these methods rely on data-independent algorithms to generate hash centers, which neglect the semantic relationships between classes and may degrade retrieval performance. This paper introduces the concept of semantic hash centers, building on the idea of traditional hash centers. We hypothesize that hash centers of semantically related classes should have closer Hamming distances, while those of unrelated classes should be more distant. To this end, we propose a three-stage framework, SHC, to generate hash codes that preserve semantic structure. First, we develop a classification network to identify semantic similarities between classes using a data-dependent similarity calculation that adapts to varying data distributions. Second, we introduce an optimization algorithm to generate semantic hash centers, preserving semantic relatedness while enforcing a minimum distance between centers to avoid excessively similar hash codes. Finally, a deep hashing network is trained using these semantic centers to convert images into binary hash codes. Experimental results on large-scale retrieval tasks across several public datasets show that SHC significantly improves retrieval performance. Specifically, SHC achieves average improvements of +7.26%, +7.62%, and +11.71% in MAP@100, MAP@1000, and MAP@ALL metrics, respectively, over state-of-the-art methods.

Problem

• Improving image retrieval via semantic hash centers
• Addressing neglect of class semantic relationships in hashing
• Enhancing hash code discriminability across diverse datasets

Innovation

• Uses semantic hash centers for image retrieval
• Generates hash codes preserving semantic structure
• Optimizes hash centers with data-dependent similarity