Application of MobileNet V2 Architecture for Chinese Chess Classification Using DNN

Abstract

Currently, there are many methods available in the use of Deep Neural Networks (DNNs), but most of them really need devices that support high-level GPUs to get maximum results, and in this study aims to get good results by using devices that do not require the use of GPUs. Therefore, in this study, MobileNetV2 is used as a model used for DNN where this model is very easy to use without having to have a GPU. And the study has results that show that the DNN model using MobileNetV2 is able to distinguish objects with an average accuracy of 98.30%, precision of 98.70%, and recall of 98.90%. These findings have the potential to have a significant impact on various fields and industries, or on the game of Chinese chess, because the dataset used in this study is Chinese chess.