tovnet-en

Violence Detection/Fall Detection

· Utilizing deep learning-based image analysis technology to detect situations such as violence and falls with high accuracy

· Violence Detection: ① Learning various key points of violence → ② Detecting violence through pose estimation on CCTV → ③ Grouping the perpetrator and victim → ④ Tracking violence to the next frame and performing real-time information processing

· Fall Detection: ① Learning various key points of fall → ② Real-time alarm when a fall occurs on CCTV

Excellence in Monitoring & Platform

+ It has automated monitoring.

· It provides information technology that detects events in real time and enables proactive responses based on an automated analysis server.

+ It resolved the staffing issue through automated day-and-night detection.

· It eliminated the need for monitoring personnel through day-and-night automated detection.

+ It solved the problem by diversifying the platform.

· It provides an automated analysis dashboard for violent incidents by time and accident type through platform development, .

· It diversifies usage environments through real-time monitoring on smartphones in addition to PCs.

Violence & Fall Detection

+ It resolved the issues with the detection method.

· A bottom-up approach that predicts keypoints in real time, analyzes their relationships, and predicts pose

+ It fixed the accuracy issue.

· Through development of a dedicated ensemble model for violence and fall detection, it enhanced accurary while maintaining fast recognition speed.

+ It resolved the problem of violence detection.

· When a violent situation occurs, it is automatically grouped and resolved as a single incident.