PERBANDINGAN CITRA RGB DAN GRAYSCALE UNTUK KLASIFIKASI WAJAH BERMASKER MENGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK

Abstract

significance of mask usage, many individuals have not been using them correctly. In order to address this issue, a system for image processing was developed using Convolutional Neural Network (CNN) capable of identifying three categories: wearing masks correctly, wearing masks incorrectly, and not wearing a mask. The dataset employed consists of 500 images for each category. Additionally, a comparison between using RGB and grayscale images was conducted, revealing that the RGB-based image model achieved an accuracy of 94.78%, while the Grayscale image reached an accuracy of 89.22%. These results confirm the importance of color information in the mask-wearing position detection process. Through this research, it is hoped to contribute to raising public awareness regarding the significance of proper mask usage in the efforts against Covid-19.