Result: Low-cost system for identification of cataract maturity using LeNet CNN.

Title:
Low-cost system for identification of cataract maturity using LeNet CNN.
Authors:
Labib, Radimas Putra Muhammad Davi1 (AUTHOR) radimas@lecturer.itn.ac.id, Syahputra, Dwangga Rizqia Meidyan1 (AUTHOR) 1912055@scholar.itn.ac.id, Maturbongs, Ririn Katherina1 (AUTHOR) 1912049@scholar.itn.ac.id, Widyatamara, Amandarika1 (AUTHOR) 1912061@scholar.itn.ac.id, Aditama, Mochamad Bayu1 (AUTHOR) 1912011@scholar.itn.ac.id, Asyifa, Elvan Dwi Nur1 (AUTHOR) 2112055@scholar.itn.ac.id
Source:
AIP Conference Proceedings. 2024, Vol. 3077 Issue 1, p1-6. 6p.
Subject Terms:
*CONVOLUTIONAL neural networks, *SYSTEM identification, *PYTHON programming language, *CATARACT, *RASPBERRY Pi, *LIBRARY technical services
Database:
Academic Search Index

Further Information

Cataracts can be removed through a surgical procedure when they reach the mature stage. Therefore, it is necessary to identify the maturity of the cataract. This paper presents the development of a cataract maturity identification system. The system is built using low-cost hardware such as Raspberry Pi 3 model B+ as the main controller and ESP32-CAM to take pictures of the eye area. The software is designed using Python programming language with OpenCV library for image processing and TensorFlow module for Convolutional Neural Network (CNN) implementation. The identification process is built using the CNN algorithm for classification with LeNet network model. The dataset consists of 165 images divided by 99 images for training purposes and 66 images for validation purposes. The results show that the designed algorithm can be embedded in low-cost devices and the identification process can be carried out with an accuracy rate of 95.45%. [ABSTRACT FROM AUTHOR]