Hospital Corridor Personnel Density Estimation Dataset

In public places like hospitals, accurately estimating personnel density can effectively enhance safety and management efficiency. However, conventional surveillance systems cannot provide precise personnel density information. Existing solutions typically rely on single-frame image analysis, lacking dynamic tracking and accurate counting. Additionally, single-angle cameras are often subject to occlusion, making it difficult to capture complete information. This dataset aims to address the issue of personnel density estimation in real-time video frames, enhancing the intelligence level of surveillance systems. The data is collected from high-definition video captured by surveillance cameras in hospital corridors. The cameras are installed at a high position to obtain the largest possible view. Regarding quality control, the data is manually annotated in multiple rounds and reviewed by industry experts to ensure annotation accuracy and consistency. The annotation team is composed of ten members with professional backgrounds in computer vision and image processing. Data preprocessing includes background modeling, denoising, frame differencing, etc., to filter noise and enhance information validity. The data is finally stored in MP4 format, organized into directories by date and camera location.