A complementary method for automated detection of microaneurysms in fluorescein angiography fundus images to assess diabetic retinopathy

PATTERN RECOGNITION

卷:46

Article

Tavakoli, Meysam^1,2,3  Shahri, Reza Pourreza^3,4  Pourreza, Hamidreza^3,5  Mehdizadeh, Alireza^6,7  Banaee, Touka^3,8  Toosi, Mohammad Hosein Bahreini⁹ 

[1] Oklahoma State Univ, Dept Phys, Stillwater, OK 74078 USA

[2] Mashhad Univ Med Sci, Ctr Res Med Phys & Biomed Imaging, Mashhad, Iran

[3] MUMS & Ferdowsi Univ Mashhad, Eye Image Anal Res Grp, Mashhad, Iran

[4] Univ Texas Dallas, Dept Elect Engn, Dallas, TX 75230 USA

[5] Ferdowsi Univ Mashhad, Dept Comp Engn, Mashhad, Iran

[6] Shiraz Univ Med Sci, Ctr Res Med Phys, Shiraz, Iran

[7] Shiraz Univ Med Sci, Biomed Engn Image Proc Lab, Shiraz, Iran

[8] Mashhad Univ Med Sci, Khatam Al Anbia Hosp, Ophthalm Res Ctr, Mashhad, Iran

[9] Mashhad Univ Med Sci, Dept Med Phys, Sch Med, Mashhad, Iran

关键词: Computer aided diagnosis;    Diabetic retinopathy;    Fluorescein angiography;    Fundus images;    Radon transform;    Microaneurysms;   

DOI  :  10.1016/j.patcog.2013.03.011

来源: Elsevier

【 摘 要 】

Early detection of microaneurysms (MAs), the first sign of Diabetic Retinopathy (DR), is an essential first step in automated detection of DR to prevent vision loss and blindness. This study presents a novel and different algorithm for automatic detection of MAs in fluorescein angiography (FA) fundus images, based on Radon transform (RT) and multi-overlapping windows. This project addresses a novel method, in detection of retinal land marks and lesions to diagnose the DR. At the first step, optic nerve head (ONH) was detected and masked. In preprocessing stage, top-hat transformation and averaging filter were applied to remove the background. In main processing section, firstly, we divided the whole preprocessed image into sub-images and then segmented and masked the vascular tree by applying RT in each sub-image. After detecting and masking retinal vessels and ONH, MAs were detected and numbered by using RT and appropriated thresholding. The results of the proposed method were evaluated on three different retinal images databases, the Mashhad Database with 120 FA fundus images, Second Local Database from Tehran with 50 FA retinal images and a part of Retinopathy Online Challenge (ROC) database with 22 images. Automated DR detection demonstrated a sensitivity and specificity of 94% and 75% for Mashhad database and 100% and 70% for the Second Local Database respectively. (C) 2013 Elsevier Ltd. All rights reserved.

【 授权许可】

Free