Anterior Chamber Angle Shape Analysis and Classification of Glaucoma in SS-OCT Images

Abstract

Optical coherence tomography is a high resolution, rapid, and noninvasive diagnostic tool for angle closure glaucoma. In this paper, we present a new strategy for the classification of the angle closure glaucoma using morphological shape analysis of the iridocorneal angle. The angle structure configuration is quantified by the following six features: (1) mean of the continuous measurement of the angle opening distance; (2) area of the trapezoidal profile of the iridocorneal angle centered at Schwalbe's line; (3) mean of the iris curvature from the extracted iris image; (4) complex shape descriptor, fractal dimension, to quantify the complexity, or changes of iridocorneal angle; (5) ellipticity moment shape descriptor; and (6) triangularity moment shape descriptor. Then, the fuzzy k nearest neighbor (fkNN) classifier is utilized for classification of angle closure glaucoma. Two hundred and sixty-four swept source optical coherence tomography (SS-OCT) images from 148 patients were analyzed in this study. From the experimental results, the fkNN reveals the best classification accuracy (99.11 ± 0.76%) and AUC (0.98 ± 0.012) with the combination of fractal dimension and biometric parameters. It showed that the proposed approach has promising potential to become a computer aided diagnostic tool for angle closure glaucoma (ACG) disease.

Figure 1

Overview of the automatic classification system for open angle and angle closure glaucoma using SS-OCT images.

Figure 2

Anterior chamber imaging showing Schwalbe's line and yellow bounding box for region of interest (ROI) for angle analysis.

Figure 3

(a) The original closed angle SS-OCT image with contaminated vertical saturation artifact; (b) detection of the artifact by searching the abrupt changes in intensities of the SS-OCT images where the vertical dotted lines denote the region of detected vertical artifact and the horizontal dotted lines denote the mean intensity of the SS-OCT images.

Figure 4

(a) Biometric parameter measurement in SS-OCT image. (b) Extracted region of iris to compute iris curvature. SL: Schwalbe's line, AT_sl: area of trapezoidal profile of iridocorneal angle centered at SL, AOD_psl: angle opening distance 500 μm posterior from SL, H _iris: the mean curvature of iris.

Figure 5

Shape analysis on the region of interest in SS-OCT image where SL is Schwalbe's line and AOD_psl is the angle opening distance 500 μm posterior from SL.

Figure 6

(a) Segmentation of anterior chamber without artifact removal; (b) segmentation of anterior chamber, the cornea, and the anterior of iris surface with artifact removal in SS-OCT image. Green region is extracted anterior chamber and red lines are edges of cornea and anterior of iris.

Figure 7

The results of segmentation and feature extractions on SS-OCT image; (a) segmentation of anterior chamber; (b) extracted iris image for curvature analysis; (c) the resulting angle profile for biometric parameter measurement; (d) region of interest for fractal dimension analysis; (e) extracted shape for moment shape analysis; and (f) fractal image of extracted region of SS-OCT image.

Figure 8

(a) Scatter plot of FD features versus mAOD for 264 images and (b) scatter plot of FD features versus AT_sl for 264 images.

Figure 9

(a) The relationship between the fuzzy strength parameter m and the classification accuracy and (b) the relationship between the fuzzy strength parameter m and the AUC value.

Figure 10

(a) Original SS-OCT image; (b) segmented binary image; (c) segmented anterior chamber; (d) extracted boundaries of lower cornea and upper iris; (e) extracted iris image.

Algorithm 1

Segmentation of anterior chamber and edge detection.