James Lowell

Optic nerve head segmentation

Reliable and efficient optic disk localization and segmentation are important tasks in automated retinal screening. General-purpose edge detection algorithms often fail to segment the optic disk due to fuzzy boundaries, inconsistent image contrast or missing edge features. This paper presents an algorithm for the localization and segmentation of the optic nerve head boundary in low-resolution images (about 20 /spl mu//pixel). Optic disk localization is achieved using specialized template matching, and segmentation by a deformable contour model. The latter uses a global elliptical model and a local deformable model with variable edge-strength dependent stiffness. The algorithm is evaluated against a randomly selected database of 100 images from a diabetic screening programme. Ten images were classified as unusable; the others were of variable quality. The localization algorithm succeeded on all bar one usable image; the contour estimation algorithm was qualitatively assessed by an ophthalmologist as having Excellent-Fair performance in 83% of cases, and performs well even on blurred images.

Measurement of retinal vessel widths from fundus images based on 2-D modeling

Changes in retinal vessel diameter are an important sign of diseases such as hypertension, arteriosclerosis and diabetes mellitus. Obtaining precise measurements of vascular widths is a critical and demanding process in automated retinal image analysis as the typical vessel is only a few pixels wide. This paper presents an algorithm to measure the vessel diameter to subpixel accuracy. The diameter measurement is based on a two-dimensional difference of Gaussian model, which is optimized to fit a two-dimensional intensity vessel segment. The performance of the method is evaluated against Brinchmann-Hansens half height, Gregsons rectangular profile and Zhous Gaussian model. Results from 100 sample profiles show that the presented algorithm is over 30% more precise than the compared techniques and is accurate to a third of a pixel.

An automated retinal image quality grading algorithm

This paper introduces an algorithm for the automated assessment of retinal fundus image quality grade. Retinal image quality grading assesses whether the quality of the image is sufficient to allow diagnostic procedures to be applied. Automated quality analysis is an important preprocessing step in algorithmic diagnosis, as it is necessary to ensure that images are sufficiently clear to allow pathologies to be visible. The algorithm is based on standard recommendations for quality analysis by human screeners, examining the clarity of retinal vessels within the macula region. An evaluation against a reference standard data-set is given; it is shown that the algorithms performance correlates closely with that of clinicians manually grading image quality.

Automated diagnosis of referable maculopathy in diabetic retinopathy screening

This paper introduces an algorithm for the automated diagnosis of referable maculopathy in retinal images for diabetic retinopathy screening. Referable maculopathy is a potentially sight-threatening condition requiring immediate referral to an ophthalmologist from the screening service, and therefore accurate referral is extremely important. The algorithm uses a pipeline of detection and filtering of “peak points” with strong local contrast, segmentation of candidate lesions, extraction of features and classification by a multilayer perceptron. The optic nerve head and fovea are detected, so that the macula region can be identified and scanned. The algorithm is assessed against a reference standard database drawn from the Birmingham City Hospital (UK) diabetic retinopathy screening programme, against two possible modes of use: independent screening, and pre-filtering to reduce human screener workload.

Quantification of Diabetic Retinopathy using Neural Networks and Sensitivity Analysis

The design of neural network classifiers for the identification of diabetic retinopathy is discussed. Red-free digitised fundal images are tiled, and a neural network is trained to distinguish exudates from drusen (similar appearing lesions). By quantifying the degree of retinopathy, the approach can be used to screen diabetic patients for referral. A novel form of hierarchical feature selection using sensitivity analysis is presented. The resulting neural network is compact, and achieves 91% sensitivity and specificity on a test set.

Assessment of stereoscopic optic disc images using an autostereoscopic screen – experimental study.

BackgroundStereoscopic assessment of the optic disc morphology is an important part of the care of patients with glaucoma. The aim of this study was to assess stereoviewing of stereoscopic optic disc images using an example of the new technology of autostereoscopic screens compared to the liquid shutter goggles.MethodsIndependent assessment of glaucomatous disc characteristics and measurement of optic disc and cup parameters whilst using either an autostereoscopic screen or liquid crystal shutter goggles synchronized with a view switching display. The main outcome measures were inter-modality agreements between the two used modalities as evaluated by the weighted kappa test and Bland Altman plots.ResultsInter-modality agreement for measuring optic disc parameters was good [Average kappa coefficient for vertical Cup/Disc ratio was 0.78 (95% CI 0.62–0.91) and 0.81 (95% CI 0.6–0.92) for observer 1 and 2 respectively]. Agreement between modalities for assessing optic disc characteristics for glaucoma on a five-point scale was very good with a kappa value of 0.97.ConclusionThis study compared two different methods of stereo viewing. The results of assessment of the different optic disc and cup parameters were comparable using an example of the newly developing autostereoscopic display technologies as compared to the shutter goggles system used. The Inter-modality agreement was high. This new technology carries potential clinical usability benefits in different areas of ophthalmic practice.