Implementasi Algoritma Convolutional Neural Network Dalam Mengklasifikasi Kesegaran Buah Berdasarkan Citra Buah

Abstract

The development of Information Technology today, which continues to grow, can help overcome various problems because matters relating to the advancement of Information Technology have spread to almost all levels of Indonesian society. Along with the development of Information Technology, it is also marked by Artificial Intelligence which can simulate human intelligence and help handle tasks in the real world. By utilizing Information Technology, one of them can be used in terms of the classification of fruit freshness. Where this classification will be very useful and help farmers and fruit consumers. This study describes the use of the Convolutional Neural Network to classify the freshness of the following fruits: apples, oranges, and bananas. And also using six classes, namely fresh apples, fresh oranges, fresh bananas, unfresh apples, unfresh oranges, and unfresh bananas. The first thing to do is Convolutional Neuronal Network training using an image dataset as input using data sources from Kaggle.com, published by "Student at Stony Brook University, New York, United States". To determine the performance of the various models produced, the following Confusion Matrix is used: accuracy, precision, and recall. The best average obtained is 93%.