Quantitative Evaluation of Hand Washing Skills Based on Convolutional Neural Network for Nursing Education

Kazushi Yamamoto¹, Miho Yoshii², Fumiya Kinoshita^1* and Hideaki Touyama¹

¹Information System Engineering, Graduate School of Engineering, Toyama Prefectural University, 5180 Kurokawa, Imizu-shi, Toyama 939-0398, Japan
²Department of Fundamental Nursing, University of Toyama, 2630 Sugitani, Toyama-shi, Toyama 930-0194, Japan
^*E-mail address: f.kinoshita@pu-toyama.ac.jp

(Received January 20, 2020; Accepted March 8, 2020)

Abstract. The World Health Organization (WHO) reports that hand hygiene education is important for health workers. However, handwashing methods in hospitals require specialized handwashing knowledge. In this paper, we present a new quantitative handwashing skills assessment method based on the Convolutional Neural Network (CNN). An image capture box with blacklights was created and photos were taken before and after handwashing after the application of fluorescent paint. The photos after handwashing were labeled by specialists in each part. The total number of photos after handwashing labeled by specialists was 585. After that, 9-fold cross-validation was performed on the photos after handwashing to calculate the discrimination accuracy (F-value). As a result, the F-value was 81.26 ± 6.18% in the handwashing image evaluation system using CNN.

Keywords: Hand Hygiene, Hand Washing Skills, Convolutional Neural Network (CNN), Classification