<![CDATA[The face is one of the distinctive characteristics of an individual and is often used to identify or distinguish one from another. The face itself has several characteristics, one of which is the shape of the face. The shape of the face has quite an important role in matters related to appearance. One example is the application in the fashion sector, where the structure of the face shape is a determining factor in choosing a hairstyle, choosing eyeglass frames, make-up and other aspects. This research focuses on comparing several types of CNN architectures such as InceptionV3, MobileNetV3, VGG-19 and CNN itself and the effect of increasing the intensity of pixel values using the histogram equalization method is also carried out. As well as implementing the system using the Flutter framework for development to the mobile platform. From the research results, it was concluded that the VGG-19 method coupled with histogram equalization succeeded in getting an accuracy level of 79.84%.]]>
Copyrights © 2023
