KnifeHunter: Structured Local Representation Learning for Fine-Grained Knife Image Retrieval in Law Enforcement

2026-08-07

AI Summary

This study addresses the pressing need for efficient, fine-grained retrieval tools in law enforcement contexts where knife-related violent incidents are increasingly common. To this end, the authors propose an end-to-end forensic knife image retrieval system that integrates global contextual and local discriminative features through a compact CoRe-Net architecture. This architecture leverages Structured Complementary Representation Learning (SCRL) and a Bidirectional Dual-level Representation Fusion (BDRF) mechanism, enhanced by residual projection and a gated local-to-global injection strategy. The work contributes the KnifeHunter dataset—comprising 25,843 images—and a standardized evaluation protocol. The system achieves 88.0% mAP under the Medium protocol and maintains robust performance with 85.1% mAP under distractor conditions. Notably, it attained 99.2% mP@1 during real-world deployment in the UK’s “Operation Sceptre.”

Abstract

Knife-enabled violence presents a major public safety challenge, and law enforcement agencies require scalable tools for catalogue-level knife identification, intelligence analysis, and source attribution. Manual visual comparison is specialist, time-consuming, and difficult to scale under operational imaging conditions. We introduce KnifeHunter, an end-to-end forensic knife image retrieval system developed with UK law enforcement. The work contributes the KnifeHunter dataset, comprising 25,843 images across 543 knife classes from police evidence, retail catalogues, and border-force seizures, with structured metadata, Medium/Hard evaluation protocols, and large-scale distractor evaluation. We further propose CoRe-Net, a compact single-descriptor retrieval architecture that combines global context with spatially localised discriminative evidence. CoRe-Net introduces Structured Complementary Representation Learning (SCRL) to organise local evidence into complementary prototype-based representations, and Bi-Directional Reciprocal Fusion (BDRF) to integrate global and local evidence through residual projection and gated local-to-global injection. Using an EVA02-Base backbone and cosine-similarity retrieval, CoRe-Net achieves 88.0% mAP and 86.7% mP@10 on the Medium protocol, and 85.1% mAP and 83.8% mP@10 under distractor conditions. KnifeHunter was deployed by UK police forces during Operation Sceptre deployments from 2023 to 2025, achieving 99.2% mP@1 on field queries. These results demonstrate a practical and effective multimedia retrieval framework for fine-grained forensic knife matching in operational law-enforcement settings.

Problem

• fine-grained retrieval
• knife identification
• law enforcement
• forensic image analysis
• scalable visual comparison

Innovation

• Structured Complementary Representation Learning
• Bi-Directional Reciprocal Fusion
• fine-grained image retrieval
• forensic multimedia analysis
• CoRe-Net