Bashir Al-Diri

An Active Contour Model for Segmenting and Measuring Retinal Vessels

This paper presents an algorithm for segmenting and measuring retinal vessels, by growing a ldquoRibbon of Twinsrdquo active contour model, which uses two pairs of contours to capture each vessel edge, while maintaining width consistency. The algorithm is initialized using a generalized morphological order filter to identify approximate vessels centerlines. Once the vessel segments are identified the network topology is determined using an implicit neural cost function to resolve junction configurations. The algorithm is robust, and can accurately locate vessel edges under difficult conditions, including noisy blurred edges, closely parallel vessels, light reflex phenomena, and very fine vessels. It yields precise vessel width measurements, with subpixel average width errors. We compare the algorithm with several benchmarks from the literature, demonstrating higher segmentation sensitivity and more accurate width measurement.

REVIEW - A reference data set for retinal vessel profiles

This paper describes REVIEW, a new retinal vessel reference dataset. This dataset includes 16 images with 193 vessel segments, demonstrating a variety of pathologies and vessel types. The vessel edges are marked by three observers using a special drawing tool. The paper also describes the algorithm used to process these segments to produce vessel profiles, against which vessel width measurement algorithms can be assessed. Recommendations are given for use of the dataset in performance assessment. REVIEW can be downloaded from http://ReviewDB.lincoln.ac.uk.

Validating retinal fundus image analysis algorithms: Issues and a proposal

This paper concerns the validation of automatic retinal image analysis (ARIA) algorithms. For reasons of space and consistency, we concentrate on the validation of algorithms processing color fundus camera images, currently the largest section of the ARIA literature. We sketch the context (imaging instruments and target tasks) of ARIA validation, summarizing the main image analysis and validation techniques. We then present a list of recommendations focusing on the creation of large repositories of test data created by international consortia, easily accessible via moderated Web sites, including multicenter annotations by multiple experts, specific to clinical tasks, and capable of running submitted software automatically on the data stored, with clear and widely agreed-on performance criteria, to provide a fair comparison.

Automated analysis of retinal vascular network connectivity

This paper describes an algorithm that forms a retinal vessel graph by analysing the potential connectivity of segmented retinal vessels. Self organizing feature maps (SOFMs) are used to model implicit cost functions for the junction geometry. The algorithm uses these cost functions to resolve the configuration of local sets of segment ends, thus determining the network connectivity. The system includes specialized algorithms to handle overlapping vessels. The algorithm is tested on junctions drawn from the public-domain DRIVE database.

Manual measurement of retinal bifurcation features

This paper introduces a new computerized tool for accurate manual measurement of features of retinal bifurcation geometry, designed for use in investigating correlations between measurement features and clinical conditions. The tool uses user-placed rectangles to measure the vessel width, and lines placed along vessel center lines to measure the angles. An analysis is presented of measurements taken from 435 bifurcations. These are compared with theoretical predictions based on optimality principles presented in the literature. The new tool shows better agreement with the theoretical predictions than a simpler manual method published in the literature, but there remains a significant discrepancy between current theory and measured geometry.

Diabetic retinopathy: current and future methods for early screening from a retinal hemodynamic and geometric approach

Diabetic retinopathy (DR) is a major disease and is the number one cause of blindness in the UK. In England alone, 4200 new cases appear every year and 1280 lead to blindness. DR is a result of diabetes mellitus, which affects the retina of the eye and specifically the vessel structure. Elevated levels of glucose cause a malfunction in the cell structure, which affects the vessel wall and, in severe conditions, leads to their breakage. Much research has been carried out on detecting the different stages of DR but not enough versatile research has been carried out on the detection of early DR before the appearance of any lesions. In this review, the authors approach the topic from the functional side of the human eye and how hemodynamic factors that are impaired by diabetes affect the vascular structure.

Joining retinal vessel segments

A new method is introduced for joining vessel segments together to form a vessel graph. Using a reference image set from the Sunderland Eye Infirmary, we analysed the retinal bifurcation geometry, to define measurements for the geometrical junction features. These distinctive measurements are employed to resolve the junctions. Self organized feature maps (SOFM) are used to ldquolearnrdquo cost functions for forming bifurcation and bridge forms. The system joins segments depending on their ldquoprojective intersectionsrdquo and the SOFM cost functions. The system includes algorithms to handle overlapping and parallel segments. Transferring the vascular network to a vascular graph provides an opportunity to extract more information and to calculate features that have been not previously calculated, by providing new measurements from graph theory.

Automatic localization of the optic disc in retinal fundus images using multiple features

Accurate optic disc localization is an essential step for a reliable retinal screening system. Existing methods for the optic disc localization may fail when encountering distractors such as imprecise boundaries, deceptive edge features and inconsistent contrast in retinal images. This paper presents an algorithm (Multi-Scheme method) for localization of the optic disc. The algorithm involves prior domain knowledge such as the optic disc size, cup-to-disc ratio (CDR) and vessel convergence feature to evaluate the confidence level for the candidate region(s) at each thresholding level. Based on the confidence level, the algorithm heuristically decides whether or not to opt for multi-scheme policy for a given image. For optimization, the Computed Response (CR) from variant versions of the same image is calculated in parallel and fits a contour to the optic disc through an iterative process of updating the location of the centre of the contour. The proposed approach has been validated using dataset ONHSD [3] and diaretdb0 [16]; and the results show the robustness and reliability of the proposed method even in the presence of distractors.

Automated Measurements of Retinal Bifurcations

This paper presents an analysis of the bifurcations of retinal vessels. The angles and relative diameters of blood vessels in 230 bifurcations were measured using a new automated procedure, and used to calculate the values of several features with known theoretical properties. The measurements are compared with predictions from theoretical models, and with manual measurements. The automated measurements agree with the theoretical prediction measurements with slightly different bias. The automated method can measure large number of retinal bifurcations very rapidly, and may be useful in correlating bifurcation geometry with clinical conditions.

A manually-labeled, artery/vein classified benchmark for the DRIVE dataset

The classification of retinal vessels into arteries and veins is an important step for the analysis of retinal vascular trees, for which the scientists have proposed several classification methods. An obvious concern regarding the strength of these methodologies is the closeness of the result of a particular method to the gold standard. Unfortunately, the research community lacks benchmarks, resulting in increased subjective error, biased opinion and an uncertain progress. This paper introduces a manually-labeled, artery/vein categorized gold standard image database, as an extension of the most widely used image set DRIVE. The labeling criterion is set after a careful analysis of the physiological facts about the retinal vascular system. In addition, the labeling process also includes several versions of original images to get certainty. A two-step validation phase consists of verification from the trained computer vision observers and a professional ophthalmologist, followed by a comparison with a gold standard set for the junction locations introduced in V4-Like filters. Our gold standard is in highly reliable form; offers research community for the result comparison and progress evaluation.