Klasifikasi Rumah Adat di Indonesia Berbasis Citra Menggunakan Convolutional Neural Network

This research focuses on developing an image classification model of traditional houses in Indonesia using Convolutional Neural Network. Indonesia is known to have a variety of traditional houses that reflect the culture and local wisdom of each region. However, this diversity is starting to be threatened by the influence of foreign cultures and the lack of interest of the younger generation. To preserve and facilitate the recognition of traditional houses, this study utilizes deep learning technology with CNN models such as VGG-16, MobileNetV2, and Xception. This study determines the best accuracy in recognizing and classifying various types of traditional houses. The dataset used consists of ten categories of traditional houses, with five additional categories to increase the variety and representation of the data. The results showed an improvement in classification accuracy of up to 87% after augmentation and hyperparameter tuning on the MobileNetV2 model.