Image Classification of Indonesian Snacks using Convolutional Neural Network

Abstract

Each region in Indonesia has its own unique and distinctive culinary traditions. However, many people are still unfamiliar with the names of traditional Indonesian snacks, especially those that originate from regions other than their own. Promoting these traditional snacks is essential as an initial step in educating both domestic and international audiences about Indonesia’s cultural diversity. Culinary heritage is also a key factor in attracting tourists to visit a region. One way to address this issue is through image classification of Indonesian traditional snacks using Convolutional Neural Networks (CNN). This study uses a dataset consisting of 30 images across 10 classes, with 3 images per class. The model was trained over 40 epochs and achieved an accuracy of 86%. The testing phase yielded a recall of 86%, precision of 91%, and an F1-score of 88%.