An Implementation of Convolutional Neural Network for Coffee Beans Quality Classification in a Mobile Information System
Robby Janandi, Tjeng Wawan Cenggoro
Abstract
Due to its massive trading in world markets, maintaining the quality of coffee is vital for the exporting countries. One approach for quality control is to have a system that can classify coffee beans based on the quality. This system can assist the small-medium coffee enterprises to monitor and secure their procurement. However, the coffee beans quality classification technology is currently unavailable to the small-medium coffee enterprises community. To address this issue, we developed a mobile application powered by a deep-learning-based model to automatically classify coffee beans quality via a mobile phone camera. The deep learning model used is chosen between ResNet-152 and VGG16 based on their performance to classify coffee beans quality. The result shows that ResNet-152 could achieve the highest accuracy of 73.3% and could also be embedded in a functional mobile application.