|
مقاله
|
Abstract
|
|
|
|
Title:
|
Automated Detection of non-Perfused Capillary in Fundus Fluorescein Angiogram of Diabetic Retinopathy Eye using a New Image processing Method
|
|
Author(s):
|
Seyed Hossein Rasta 1 , Shima Nikfarjam 1 , Alireza Javadzadeh 2, Hadi Seyedarabi 3
|
|
Presentation Type:
|
Oral
|
|
Subject:
|
Posterior Segment
|
|
Others:
|
|
|
Presenting Author:
|
|
|
Name:
|
Seyed Hosssein Rasta
|
|
Affiliation :(optional)
|
1 Dept. of Medical physics,/ 2. Dept. of Ophthalmology, Tabriz University of Medical Sciences, Iran. , 3 Faculty of Electrical and Computer Eng., University of Tabriz, Iran
|
|
E mail:
|
s.h.rasta@abdn.ac.uk
|
|
Phone:
|
|
|
Mobile:
|
09370630833
|
|
|
|
|
Purpose:
|
This paper presents a new method for automated detection of non-perfused (NP) regions on fundus fluorescein angiogram (FFA).
|
|
Methods:
|
Capillary Non-perfusion (CNP) is one of the retinal vascular diseases in diabetic retinopathy (DR) patients usually diagnosed with diabetes mellitus. Whilst major vessels appear as ridges, non-perfused areas are usually observed as ponds that are surrounded by healthy capillaries in FA images. These lesions were found as smoothed regions with very high uniformity, low entropy, and small intensity variations in FFA images. The strategy we have adopted is screening the whole image with a fixed window size, which is small enough to enclose areas with identified topographic characteristics. In order to reject false candidates, we also perform a thresholding operation on the screen and marked images. The proposed detection algorithm has been tested and validated on 541 FFA images and has also been certified by an ophthalmologist.
|
|
Results:
|
The results showed within 81% sensitivity and 78% specificity from the ideal experts’ opinion.
|
|
Conclusion:
|
This method could enhance an automated diagnostic system to detect non-perfusion lesions in FFA images showing diabetic retinopathy or referable diabetic retinopathy.
|
|
Attachment:
|
|
|