Pemanfaatan algoritma Convolutional Neural Network (CNN) untuk klasifikasi jenis noken

Abstrak

Noken is a traditional bag from Papua that holds great cultural value and has been recognized by UNESCO as Intangible Cultural Heritage. The diversity of noken types based on motifs, shapes, and regions of origin presents a challenge for the identification process, which is still carried out manually. This study aims to develop an automatic noken image classification system using a Convolutional Neural Network (CNN) with transfer learning. Three CNN architectures used in this study are VGG16, InceptionV3, and MobileNetV2. The dataset consists of 500 noken images, divided into two types: Bitu Agia and Junum Ese. The training process was conducted using the TensorFlow library with the best parameters. Evaluation was carried out using accuracy, precision, recall, and F1-score metrics, as well as graphical visualizations of accuracy and loss. The results showed that MobileNetV2 achieved the best performance with an accuracy of 97 percent, followed by InceptionV3 with 96 percent, and VGG16 with 87 percent. This study demonstrates that the deep learning approach is effective in the image classification of cultural objects and can support the digital preservation of Papuan culture.