RFHNet: Relational and Frequency-Aware Hashing Network for Large-Scale Fine-Grained Food Image Retrieval

2026-07-07

AI Summary

Existing hashing-based retrieval methods struggle to effectively capture local semantic cues and frequency-domain sensitive features in fine-grained food images, limiting their retrieval performance. To address this challenge, this work proposes RFHNet, a cascaded hierarchical hashing network that jointly optimizes multi-level representations of both global structure and local details through three core components: Fine-grained Relation Modeling (FRM), Multi-frequency Modulation Fusion (MFMF), and Hierarchical Semantic Synergy (HSS). Extensive experiments on six food-specific benchmarks demonstrate that RFHNet significantly outperforms state-of-the-art methods, achieving mAP improvements of 4.44%–17.20% with 12-bit hash codes, thereby confirming its practical utility in applications such as intelligent food services.

Abstract

Fine-grained food image retrieval is a key task in computational gastronomy, with applications in food traceability, dietary monitoring, and smart catering systems. Although hashing-based retrieval is attractive for large-scale search due to its storage efficiency and fast Hamming-distance computation, existing methods often perform poorly in fine-grained food scenarios, where subtle local semantics and frequency-sensitive visual cues are essential. To address this challenge, we propose RFHNet, a cascaded hierarchical hashing network that captures both global structure and fine-grained local details through multi-level representations. RFHNet includes three components: (1) Fine-grained Relation Modeling (FRM) to capture subtle visual differences among similar food components; (2) Multi-Frequency Modulated Fusion (MFMF) to extract informative multi-frequency features; and (3) Hierarchical Semantic Synergy (HSS) to adaptively integrate multi-level representations and generate discriminative hash codes. Experiments on six food-specific benchmarks show that RFHNet consistently outperforms state-of-the-art hashing methods, with mAP gains of 4.44\% to 17.20\% at 12 bits. These results validate the effectiveness of RFHNet for large-scale visual food retrieval and smart catering applications. The source code will be released upon publication.

Problem

• fine-grained food image retrieval
• hashing-based retrieval
• local semantics
• frequency-sensitive visual cues
• large-scale retrieval

Innovation

• Fine-grained Relation Modeling
• Multi-Frequency Modulated Fusion
• Hierarchical Semantic Synergy
• Hashing Network
• Food Image Retrieval