MA-CBP: A Criminal Behavior Prediction Framework Based on Multi-Agent Asynchronous Collaboration

TL;DR

This paper introduces MA-CBP, a novel multi-agent asynchronous framework that improves real-time criminal behavior prediction by integrating semantic video analysis and historical context reasoning, enhancing urban safety monitoring.

Contribution

The paper presents a new multi-agent asynchronous collaboration framework for criminal behavior prediction, incorporating semantic video understanding and multi-scale annotations, outperforming existing methods.

Findings

Achieves superior performance on multiple datasets.

Enables early warning of criminal activities.

Provides a high-quality, multi-scale annotated criminal behavior dataset.

Abstract

With the acceleration of urbanization, criminal behavior in public scenes poses an increasingly serious threat to social security. Traditional anomaly detection methods based on feature recognition struggle to capture high-level behavioral semantics from historical information, while generative approaches based on Large Language Models (LLMs) often fail to meet real-time requirements. To address these challenges, we propose MA-CBP, a criminal behavior prediction framework based on multi-agent asynchronous collaboration. This framework transforms real-time video streams into frame-level semantic descriptions, constructs causally consistent historical summaries, and fuses adjacent image frames to perform joint reasoning over long- and short-term contexts. The resulting behavioral decisions include key elements such as event subjects, locations, and causes, enabling early warning of potential criminal activity. In addition, we construct a high-quality criminal behavior dataset that provides multi-scale language supervision, including frame-level, summary-level, and event-level semantic annotations. Experimental results demonstrate that our method achieves superior performance on multiple datasets and offers a promising solution for risk warning in urban public safety scenarios.