Fully Automated PCO Quantification System

Pourshahabi, Muhammad Reza

Pourreza, Hamid Reza

Daneshvar, Ramin

Nowadays, Posterior Capsule Opacification (PCO) is a common postoperative complication of cataract surgery. The rate of incidence and the intensity of PCO are affected by factors such as type and shape of implanted intraocular lens (IOL), cataract surgical techniques and etc. Clinical quantification of PCO is so subjective that evaluating the effects of these factors on PCO are varying among studies. The need for a reliable and efficient automated PCO quantification system is highly desired.

In this thesis, a new fully automated contourlet based PCO quantification system is presented using fuzzy and non-fuzzy approaches. For every PCO images, ROI is detected and then contourlet transform has been used for feature extraction stage. Multiresolution, localization, critical sampling, directionality and anisotropy are five important features for an effective image representation tool, that all of them are satisfied using contourlet transform. Contourlet transform is a double filter bank named pyramidal directional filter bank (PDFB). Contourlet transform consists of Laplacian Pyramid (LP) followed by a Directional Filter Bank. Smooth contours can be captured efficiently using contourlet transform. With consideration of the type of PCO images that contain a lot of contours, contourlet transform for feature extraction stage has been used.

Fuzzy and non-fuzzy approaches have been used in order to determine PCO score. Proposed systems have been compared with two standard subjective methods and three famous semi-automated methods. Experimental results show the reliability and correctness of both fuzzy and non-fuzzy systems.

Evaluation of the Ability of Image Processing Software of Mashhad University in Detecting the Changes in Retinal Vessel Diameter in Different Stage of Diabetic Retinopathy

Hajipoor Kh., Maryam

Banaee, Touka

Pourreza, Hamid-Reza

Propose to describe the relationship of retinal vascular caliber with diabetes and diabetic retinopathy we designed this population-based cross-sectional analysis in 100 diabetic patients in Khatamolanbia Hospital in Mashhad.

Method: Retinal photographs of 7 standard fields were taken and retinal vascular caliber was measured and summarized by using a validated computer program that design in Ferdowsi University of Mashhad that measure retinal vascular size with pixel.Results: After analysis of our data and controlling for age, gender, blood pressure, and other factors, there was a significant relationship between retinal vascular caliber and diabetic retinopathy stages, this relationship was significant in 7, 9, 11 pixels, although there was a direct relationship between age and diabetic retinopathy stages in all pixels.

Conclusions: Increasing severity of diabetic retinopathy in persons with diabetics is associated with widening of retinal vascular caliber. Emerging evidence suggests that quantitative assessment of retinal vascular caliber may provide prognostic information regarding the risk of diabetic micro-vascular complications such as retinopathy. Wider retinal vessels have been shown to predict the incidence and progression of diabetic retinopathy, in keeping with experimental work that indicates a role for retinal vessel dilation in the pathogenesis of diabetic retinopathy retinal vessel dilation cause a rise in capillary pressure, which in turn can lead to capillary wall dilation (micro-aneurysm), leakage (edema), and rupture (hemorrhage), all classical signs of diabetic retinopathy. However, not all studies have reported an association of retinal vessels dilation with risk of retinopathy. Reason for these inconsistencies are not apparent but could be related to imprecise diagnoses of diabetes. Additional studies are clearly needed.

Automated Detection of Diabetic Retinopathy Vascular Lesions (Microaneurysms) in fluorescein angiography by image processing methods

M. H. Bahreyni Toossi

Pourreza, Hamid Reza

Banaee, touka

The main objective of this research is to aid in developing automatic screening systems for DR. Such systems will significantly help ophthalmologists while diagnosing and treating patients. Automated screening systems promise with more efficient and costless medical services, in addition to delivering health-care services to rural areas.

Over and above, automated screening systems will lend a hand to hold back the personal and social costs of DR as one of the most prevalent complications of diabetes and one of the leading causes of blindness.
This thesis reports on research into an aspect of automatic grading of diabetic retinopathy; namely the identification of microaneurysms (MAs) and hemorrhages (HEs) in retinal fluorescein angiography (FA) image.