Implementasi Machine Learning untuk Mendeteksi Penyakit Katarak menggunakan Kombinasi Ekstraksi Fitur dan Neural Network Berdasarkan Citra

Abstract

According to data from the World Health Organization (WHO), more than 1.3 billion people worldwide experience visual impairments, with Cataracts being one of the main causes. Cataracts are an eye condition characterized by clouding of the lens, which can lead to blindness if left untreated. This study aims to accurately detect Cataracts using a combination of feature extraction and neural networks, utilizing digital fundus images. The Dataset used consists of 600 fundus images divided into 80% for training and 20% for testing. The feature extraction process is performed to identify distinctive characteristics of the images relevant to Cataract diagnosis. These features are then analyzed by a neural network to recognize patterns indicative of Cataracts. To optimize performance, this study implements a hypertuning process. Before tuning, the initial model achieved an accuracy of 0.83, with precision, recall, F1-score of 0.83, and an AUC of 0.92. After four stages of hypertuning, the model’s performance improved progressively. The first tuning achieved an accuracy of 0.85, with precision, recall, and F1-score of 0.85, and an AUC of 0.93. In the second tuning, accuracy increased to 0.88, with precision of 0.87, recall of 0.88, F1-score of 0.87, and an AUC of 0.93. The third tuning maintained an accuracy of 0.88, with precision improving to 0.90, recall at 0.87, F1-score of 0.88, and an AUC of 0.94. The fourth tuning delivered the best results, with an accuracy of 0.90, precision of 0.92, recall of 0.89, F1-score of 0.90, and an AUC of 0.94. These results demonstrate that the hypertuning process plays a significant role in improving model performance.