Abstract
This study designed an application for the recognition of early kidney disorders through digital images of the iris using the convolutional neural network (CNN) method with the Raspberry Pi 3 model B + interface. The best accuracy results obtained by varying the number of epochs, the learning rate value, kernel size, database composition, and the pooling layer function are 94% at epoch 12, 92% at 0.0001, 95% at 3×3, 95% on composition 100 train and 50 validation, 90% use the max pooling function.
Agustian, Indra and Faisal, Hadi and Khairul, Amri (2019) PRE-DIAGNOSIS GANGGUAN GINJAL MELALUI CITRA IRIS MATA MENGGUNAKAN RASPBERRY PI DENGAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN). AMPLIFIER Jurnal Ilmiah BidangTeknik Elektro dan Komputer, 9 (1). pp. 15-23. ISSN 2089-2020
