A novel web-based depth video rewind approach toward fall preventive interventions in hospitals

Area of Science:

• Medical Informatics
• Patient Safety Technology
• Computer Vision in Healthcare

Background:

• Patient falls in hospitals are a significant concern, leading to injury and increased healthcare costs.
• Existing fall detection systems generate alarms but lack efficient tools for post-fall investigation.
• Reviewing video footage is crucial for understanding fall mechanisms and implementing preventative strategies.

Purpose of the Study:

• To develop and implement a web-based application for reviewing depth videos of hospital rooms.
• To facilitate the investigation of event chains leading to patient falls using privacy-preserving Kinect depth imaging.
• To enhance the process of identifying fall risks and causes through novel visualization techniques.

Main Methods:

• Utilized Kinect depth sensors to capture anonymized, shadow-like video data of patients and their environment.
• Integrated a previously developed fall detection system that generates real-time fall alarms.
• Developed a novel web application featuring advanced visualization tools for efficient video search and review.

Main Results:

• Successfully implemented a functional web application enabling rewind and review of depth videos.
• The application aids in investigating the specific events and circumstances preceding patient falls.
• Visualization techniques highlight frames with potential fall risks, streamlining the investigative process for healthcare staff.

Conclusions:

• The developed web application significantly eases the review of patient fall videos, improving investigative efficiency.
• Privacy-preserving depth imaging effectively addresses patient confidentiality concerns during video surveillance.
• This tool supports healthcare providers in understanding fall etiologies, ultimately aiming to enhance patient safety and reduce fall recurrence.