Segmentasi Pembuluh Darah Retina menggunakan Fuzzy Entropy dengan Perbaikan Deteksi Pembuluh Darah Tipis

Farosanti, Lafnidita (2019) Segmentasi Pembuluh Darah Retina menggunakan Fuzzy Entropy dengan Perbaikan Deteksi Pembuluh Darah Tipis. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 05111650010005-Master_Thesis.pdf Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Diabetic Retinopathy merupakan kelainan pembuluh darah retina mata yang diakibatkan oleh komplikasi penyakit diabetes. Deteksi penyakit Diabetic Retinopathy lebih dini diperlukan agar kelainan ini dapat ditangani secara cepat dan tepat. Kelainan ini ditandai dengan melemahnya bagian pembuluh darah tipis akibat tersumbatnya aliran darah kemudian menyebabkan bengkak pada mata bahkan kebutaan. Oleh karena itu diperlukan metode analisa pembuluh darah retina melalui proses segmentasi pembuluh darah terutama pada bagian penting yaitu pembuluh darah tipis. Peneliti melakukan perbaikan deteksi pembuluh darah tipis atau Thin Vessel Enhancement untuk mendapatkan citra dengan ekstraksi pembuluh darah tipis lebih banyak. Metode ini merupakan konvolusi dengan 2 kernel yaitu a bank of line operator dan a bank of edge filter. Setelah itu dilakukan proses segmentasi menggunakan metode Fuzzy Entropy. Penggabungan metode ini menghasilkan citra segmentasi pembuluh darah dengan perbaikan pembuluh darah tipis. Segmentasi yang dihasilkan dibagi menjadi 3 kategori yaitu pembuluh darah utama, medium, dan tipis berdasarkan nilai threshold yang diperoleh dari entropy masing-masing membership function pada Fuzzy Entropy. Uji coba dilakukan terhadap metode Thin Vessel Enhancement menggunakan 1 kernel dan Fuzzy Entropy dari nilai threshold ke-1 maka diperoleh nilai accuracy, sensitivity, dan specivicity sebesar 94.81%, 66.83%, dan 97.51%.
===================================================================================================================================
Diabetic Retinopathy is a disorder of the retinal blood vessels in the eye caused by complications from diabetes. Early detection of Diabetic Retinopathy is needed so that this disorder can be dealt with quickly and precisely. This disorder is characterized by a weakening of the thin blood vessels due to blockage of the blood flow and then causing swelling in the eyes and even blindness. Therefore we need a method of analyzing retinal blood vessels through the process of segmenting blood vessels, especially in important parts, namely thin blood vessels. Researchers use thin vascular repair or Thin Vessel Enhancement to obtain images with more thin blood vessel extraction. This method is a convolution with 2 kernels, namely a bank of line operators and a bank of edge filter. After that, the segmentation process is carried out using the Fuzzy Entropy method. Combining this method produces blood vessel segmentation images with thin vessel enhancement. The resulting segmentation is divided into 3 categories, namely the main, medium and thin blood vessels based on the threshold value obtained from the entropy of each membership function on Fuzzy Entropy. Experiment was conducted on the Thin Vessel Enhancement method using 1 kernel and Fuzzy Entropy from first threshold value then obtained the value of accuracy, sensitivity, and specivicity of 94.81%, 66.83%, and 97.51%.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Diabetic Retinopathy, Thin Vessel Enhancement, Fuzzy Entropy.
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing.
Divisions:	Faculty of Information and Communication Technology > Informatics > 55101-(S2) Master Thesis
Depositing User:	Lafnidita Farosanti
Date Deposited:	23 Jul 2026 08:04
Last Modified:	23 Jul 2026 08:04
URI:	http://repository.its.ac.id/id/eprint/67917

Actions (login required)

View Item