Analysis of Unclean Hand System Detection Using Template Matching Technique

Abstract

The aim of this project is to audit the handwashing technique of hospital staff that may cause infection to the patients. This project is to detect unclean washed hands using image processing technique specifically template matching. The detection and recognition of palm in images is the key methodology of this paper. The prototype used for capturing hand images is a dark box with UV light and a camera. Target will need to apply Glogerm on their hands that imitate bacteria. Hence, when they wash their hands inappropriately, Glogerm can be seen in the captured images under the UV light as the unwanted stain on washed hands, the target handwashing technique needs to be improved. Templates of the missed area of washed hands are used to compare the correctness of hand washed techniques by the target. Data of 100 images were taken, results are; 100% accuracy of the hand image without Glogerm, 56.67% of the image that did not wash using water after applying the Glogerm and 45.45% accurate when user wash their hand by using water after applying Glogerm. The overall efficiency of the system in detecting the missed part is 51% accuracy As a summary, this project accurately detects stain percentage that represents the missed part when applying the template matching technique.