CLASSIFICATION OF BANANA RIPENESS DEGREE USING YOLO-V8

ABSTRACT The ripeness level of bananas is a key factor for both producers and consumers, affecting the entire logistics chain of the fruit and its final price. Artificial intelligence techniques can be applied to enhance the efficiency of these processes. This article presents BananaRipe, a web application with mobile-friendly computer vision features designed to determine the ripeness level of bananas using a YOLO-v8-based classification model. In particular, a web application was developed in Python using the Flask microframework, integrated with a classification model trained on a dataset containing eight banana-ripeness classes. Test results showed that BananaRipe achieved 95.96% accuracy and 97.59% recall, demonstrating highly satisfactory compared to other models in the literature.