Researchers have developed a new system for detecting moderate physical violence, such as pushing, in real-time using surveillance footage. The system employs YOLO11 and YOLO11-Pose for human detection and keypoint extraction, then uses a Random Forest classifier to analyze body inclination and joint angles to differentiate aggressive behavior. This approach achieved a precision of 0.98 in controlled settings and 0.72 in challenging real-world scenarios, demonstrating its potential for early intervention in urban security. AI
Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →
IMPACT This research could enhance public safety by enabling early detection of aggressive behavior in surveillance systems.
RANK_REASON This is a research paper published on arXiv detailing a new method for violence detection.