CCTV Incident is an open synthetic dataset specifically designed for fall detection, pose estimation, and accident monitoring in computer vision tasks. It is designed to analyze from a CCTV overhead view, enabling models to understand human postures and accurately distinguish between individuals who are standing and those who have fallen.

Dataset composition:

• It provides dual annotations, including bounding boxes and a 17-point keyframe skeleton conforming to the COCO standard layout.
• The target scenario focuses on accidents (approximately 951 TP3T, with the main subjects being lying down/falling down), making it suitable for training anomaly detection models.
• The dataset is fully synthesized, ensuring privacy and security, and posing no GDPR/privacy risks.
• Compatible with YOLOv8-Pose and YOLO11-Pose models

Data format:

• Image files are stored in the images/ directory and are high-quality composite .jpg/.png files.
• Label files are stored in the labels/ directory, are .txt files, and contain category IDs, bounding boxes, and key point information.