DESIGN OF WEBSITE BASED BACK-END APPLICATION SYSTEM FOR TRAINING IN DETECTING ABNORMALITIES OF FUNDUS IMAGES USING MANUAL ANNOTATION METHOD

• Main Author: Pratama, Aditya
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/42514
• Institution: Institut Teknologi Bandung

Diabetes mellitus (DM) is a chronic degenerative disease with the highest morbidity and mortality rates in the world. World Health Organization (WHO) reported in the year 2000 that Indonesia is the fourth most DM country in the world, with a total of 8.4 million sufferers and will be reach 21.3 million by 2030. It takes many competent ophthalmologists to continue monitoring to prevent worse complications of DR disease. Therefore, we need a platform that can help prospective ophthalmologists to practice independently in diagnosing abnormalities in the eye, especially in people with DM. One solution for practicing diagnosis is to utilize digital fundus images. In this final project, a web application called Optan will be developed as a training tool for doctors to annotate fundal image abnormalities. The application platform created must be able to provide a quantitative assessment of two different fundal image annotation results. In this final project, we will discuss the back-end system of the website annotations that have been made. The system is implemented using PHP as the main programming language consisting of subsystems: login registers, add local files, add new users, choose references, annotation to images, permit the register and check percentage of similarity. The results of the back-end system that have been made, can do all the needs on the front-end so that annotations can be made by users and quantitative annotation similarity analysis through the calculation of a jaccard index can be generated. The results of comparison of annotations conducted by 6 physician respondents showed that annotations using the circle annotation tool had the highest percentage of similarity compared to using point and line annotation tools.