Effectiveness in Preventing the Elopement Using a Face Recognition System for Surveillance Camera Images

Leaving the hospital without permission is an incident that can lead to a violation of patient safety. This study aimed to enhance precision by installing surveillance cameras at two locations: the hospital ward and the entrance. Thirty-five patients admitted to the hospital were registered for face recognition, and face detection was performed. We compared the precision rate between using only the surveillance camera images of the hospital entrance, and when the surveillance camera images of the hospital ward and the entrance were used. The results showed that the precision rate using only the hospital entrance was 18.1%. This lower rate was due to the lack of optimization for the environment. In contrast, the precision rate combining the hospital ward and the entrance reached 81.0%, demonstrating the proposed method's effectiveness in reducing false positives.