Text
Deteksi Eksudat Pada Citra Fundus Retina Menggunakan Transformasi Bottom-Hat Top-Hat dan Klasifikasi Extreme Learning Machine
ABSTRACT
The application of IT in hospitals is needed to optimize the employee performance in order
to facilitate services and provide comfort for patients. One of them is Ciremai Hospital which
has implemented IT using the Sistem Informasi Manajemen Rumah Sakit (SIMRS). To help
realize the goals and achieve the IT goals of SIMRS, it is necessary to conduct a governance
audit using the COBIT 5 framework. The respective domains used are APO1, APO7, EDM4,
DSS6, and APO6. The results of capability level for all domains is level 3 and the target to be
achieved is level 4. From the domains used will be made a recommendation based on the
domains process capability level to help achieve the target level of ability that want to be
achieved.
Keywords : IT Governance, COBIT 5, SIMRS, Capability Level.
ABSTRACT
Diabetes Melitus (DM) could lead complication to other organs, such as eye’s retina this
disease called Diabetic Retinopathy (DR). Diabetic Retinopathy is caused by damaged micro
vascular in retina. One of the characteristic that can determinant people with DR is the
emergence of exudates, which is a fluid with lipid and protein substance that comes out from
damaged micro vascular. Characteristic of exudate are bright yellowish color, size and
appearance location may varies. Segmentation are done by using Bottom-Hat and Top-Hat
transformation to seperate exudate candidates from retinal fundus images. Some features are
used to distinguish exudate and normal candidate, such as average, standard deviation,
centroid, area, and solidity are extracted from CIELUV color. Feature that obtained from
segmentation are classified by machine learning. Extreme Learning Machine (ELM) is one
of Artificial Neural Network (ANN) method that can be used to classified exudate candidate.
In this experiment ELM with sigmoid activation function can achieved specificity 95.33%,
accuracy 94.33%, and sensitivity 94%.
Keywords: ELM, Bottom-Hat, Top-Hat, Diabetic Retinopathy, Exudate Detection,
CIELUV color space.
881F20IV | 881 F 20-iv | Perpustakaan FSM Undip (Referensi) | Tersedia |
Tidak tersedia versi lain