TACR-YOLO: A Real-time Detection Framework for Abnormal Human Behaviors Enhanced with Coordinate and Task-Aware Representations

TL;DR

TACR-YOLO is a real-time detection framework that improves abnormal human behavior detection by enhancing small object recognition, resolving task conflicts, and refining multi-scale fusion, achieving high accuracy on a new dataset.

Contribution

The paper introduces TACR-YOLO, a novel real-time framework with specialized modules and optimized anchor boxes for improved abnormal behavior detection.

Findings

Achieves 91.92% mAP on PABD dataset

Outperforms existing methods in speed and robustness

Ablation studies confirm effectiveness of each module

Abstract

Abnormal Human Behavior Detection (AHBD) under special scenarios is becoming increasingly crucial. While YOLO-based detection methods excel in real-time tasks, they remain hindered by challenges including small objects, task conflicts, and multi-scale fusion in AHBD. To tackle them, we propose TACR-YOLO, a new real-time framework for AHBD. We introduce a Coordinate Attention Module to enhance small object detection, a Task-Aware Attention Module to deal with classification-regression conflicts, and a Strengthen Neck Network for refined multi-scale fusion, respectively. In addition, we optimize Anchor Box sizes using K-means clustering and deploy DIoU-Loss to improve bounding box regression. The Personnel Anomalous Behavior Detection (PABD) dataset, which includes 8,529 samples across four behavior categories, is also presented. Extensive experimental results indicate that TACR-YOLO achieves 91.92% mAP on PABD, with competitive speed and robustness. Ablation studies highlight the contribution of each improvement. This work provides new insights for abnormal behavior detection under special scenarios, advancing its progress.