Agnestasia, Ekky Melynda (2017) Deteksi Penyakit Diabetes Makula Edema pada Citra Fundus Retina menggunakan Operasi Morfologi dan Transformasi Watershed. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
5113100093-Undergraduate_Theses.pdf - Published Version Download (2MB) | Preview |
Abstract
Penyakit diabetes makula edema merupakan komplikasi dari diabetes retinopati yang dapat menyebabkan kehilangan penglihatan dan kebutaan. Penyakit diabetes makula edema didiagnosis ketika eksudat keras berada di dalam wilayah makula. Eksudat keras merupakan lipid yang berwarna kekuningan pada citra. Sedangkan makula bertanggung jawab sebagai fotoreseptor dalam penglihatan yang memiliki bentuk seperti daerah gelap.
Pada tugas akhir ini akan dilakukan pengembangan sistem deteksi penyakit diabetes makula edema pada citra fundus retina menggunakan operasi morfologi dan transformasi watershed. Dalam proses pengembangan sistem ini diperlukan tiga tahapan proses yaitu preprocessing, segmentasi dan menghitung jarak terdekat antara makula dan eksudat keras. Proses preprocessing dilakukan sebelum citra masuk ke proses segmentasi. Proses segmentasi dibagi menjadi tiga proses yaitu segmentasi optic disc, makula dan eksudat keras. Proses menghitung jarak terdekat antara makula dan eksudat keras dilakukan untuk deteksi penyakit diabetes makula edema berdasarkan Early Treatment Diabetic Retinopathy Study (ETDRS). Data masukan yang digunakan yaitu citra fundus retina yang diambil dari Methods to evaluate segmentation and indexing techniques in the field of retinal ophthalmology (MESSIDOR).
Uji coba menggunakan 30 citra menunjukkan bahwa metode ini dapat memberikan hasil deteksi yang optimal dengan accuracy sebesar 96,67%. Untuk sensitivity hasil deteksi normal, diabetes makula edema tahap 1 dan diabetes makula edema tahap 2 adalah 100%, 90% dan 100%. Sedangkan untuk specifity hasil deteksi normal, diabetes makula edema tahap 1 dan diabetes makula edema tahap 2 adalah 95%, 100% dan 100%.
===================================================================================
Diabetic macular edema is a complication of diabetic retinopathy. It can cause vision loss and blindness. Diabetic macular edema is diagnosed when the hard exudate is in the macular region. Hard exudates are lipids that color is yellow in the image. While the macula is responsible as photoreceptors in vision that have a dark areas.
In this undergraduate thesis, we propose a detection system of diabetic macular edema in a retinal fundus image using morphological operation and watershed transformation. In the process of developing this system required three stages of the process. Processes are preprocessing, segmentation and calculating the closest distance between macula and hard exudate. Preprocessing process is done before the image goes into the segmentation process. The segmentation process is divided into three processes namely optical disc segmentation, macula segmentation and hard exudate segmentation. The process of calculating the closest distance between the macula and hard exudate is performed for detection of diabetic macular edema based on the Early Treatment Diabetic Retinopathy Study (ETDRS). The input data used are retinal fundus images taken from Methods to evaluate segmentation and indexing techniques in the field of retinal ophthalmology (MESSIDOR).
Testing using 30 images shows that this method can provide optimal detection results with the accuracy of 96.67%. For the sensitivity, detection results of normal, stage 1 and stage 2 are 100%, 90% and 100%. For the specificity, detection results of normal, stage 1 and stage 2 are 95%, 100% and 100%.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | Citra Fundus Retina; Diabetes Makula Edema; Operasi Morfologi; Transformasi Watershed |
| Subjects: | R Medicine > RE Ophthalmology T Technology > T Technology (General) |
| Divisions: | Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
| Depositing User: | AGNESTASIA EKKY MELYNDA |
| Date Deposited: | 09 Aug 2017 07:39 |
| Last Modified: | 05 Mar 2019 03:35 |
| URI: | http://repository.its.ac.id/id/eprint/42278 |
Actions (login required)
 |
View Item |