Aplikasi Deteksi Sample/Citra Darah Menggunakan Metode Caps Neural Network Berbasis Android

Abstract

Information technology has now become a necessity that cannot be separated from the process of life, its development is currently very much needed in solving various problems that arise in people's lives. The bite of an infected female Anopheles mosquito is the vector for transmitting this disease. Parasites attack red blood cells after invading the body through mosquito bites and eventually settle in the liver. Innovation in healthrelated technology has developed rapidly to date. One of them is the blood checking process to find out malaria parasites. The use of various methods can certainly determine the progress of the image processing process, one image processing method other than those mentioned above is the Neural Network method. One of these artificial models is Artificial Neural Networks (ANN), which continually attempts to imitate the way the brain actually learns. The author uses research methods with literature studies collected from various sources such as books, journals, reports and other literature reviews related to research. This research focuses on developing a malaria parasite detection application using Android-based Caps Neural Network (CapsNet) for blood images. The results of this research show that the blood image sample detection recognition system using the caps method and the Convolutional Neural Network algorithm obtained the best accuracy results of 96.92%. This level of accuracy is influenced by the learning rate value in the training process, apart from that the level of blood image resolution, the amount of training data and test data and the number of layers in the CNN architecture also have an influence.