Classification Of the Maturity Level of Fermented Glutinous Rice Using Convolutional Neural Network Model

Fermented glutinous rice is a popular snack in Indonesia. One of the main benefits of eating white cheddar rice is to trigger the digestive system. Excessive consumption can result in a decrease in sweetness and inappropriate texture. Therefore, it is necessary to classify the maturity level of the tape, so that there is no excessive maturity that results in adverse effects on the body and the quality of the tapes. The study aims to test the accuracy of the white tape maturity classification program as well as design and implement a classification system using the Convolutional Neural Network (CNN) method with the VGG16 architecture. The white tape image data set was obtained with the iPhone X camera in jpg format, covering three maturity classes: raw, ripe, and rotten, each consisting of 400 images. The data set is divided into 768 training data, 192 validation data, and 240 test data, then processed through preprocessing stages including resize, augmentation, and rescale. The CNN model was implemented with the VGG16 architecture and tested on various Epochs, producing an accuracy of 0.98 on Epochs 20 and 30, and reaching 0.99 on the 40th. The results of the research showed that the CNN method with VGG-16 architecture was effective in classifying the maturity level of the tape, achieving high accuration and significant consistency as the number of Epochs increased. This implementation is expected to preserve the quality of the tapes and extend the application of modern technology in traditional industries