Blood vessels extraction from fundus fluorescein angiogram in Curvelet domain
Autor: | Amir Ali Amini Tehrani, Hossein Ebrahimpour-Komleh, Davood Aghadoost |
---|---|
Rok vydání: | 2018 |
Předmět: |
medicine.diagnostic_test
business.industry Computer science 010103 numerical & computational mathematics 02 engineering and technology Filter (signal processing) Fundus (eye) Fluorescein angiography 01 natural sciences Edge detection Image (mathematics) Hough transform law.invention law 0202 electrical engineering electronic engineering information engineering Curvelet medicine 020201 artificial intelligence & image processing Segmentation Computer vision Artificial intelligence 0101 mathematics business |
Zdroj: | 2018 6th Iranian Joint Congress on Fuzzy and Intelligent Systems (CFIS). |
DOI: | 10.1109/cfis.2018.8336631 |
Popis: | This paper proposes a new method for blood vessel segmentation and extraction based on Curvelet domain and Hough transform. Diabetic retiopathy is an important human visual disease which leads blindness. Fluorescein angiography has become indispensable in the diagnosis and evaluation of many retinal conditions. Here, we introduced a novel and efficient algorithm for automated retinal blood vessels segmentation in fluorescein angiography images. The proposed method in this paper consists of three main steps that in the first step, the input image to create an image with high contrast and uniform background by becoming curvelet transform under processing. Then, in the next step Kirschs templates are used for edge detection in retinal images and arithmetic mean filter is applied to the image. Finally, Hough transform is applied to the image obtained from the previous step to remove non-information parts. Experimental results show the superiority of the proposed method against competing approaches. |
Databáze: | OpenAIRE |
Externí odkaz: |