Image analysis algorithm for the Anterior Chamber Angle Closure estimation and Van Herick classification

Abstract

Screening activity is essential for the prevention of diseases such as glaucoma. Concerning primary angle closure glaucoma, the anterior chamber angle can be monitored to evaluate the disease’s progress. Van Herick technique is a non-invasive qualitative approach for estimating the angle aperture. In our previous papers, we presented an automatic instrument able to perform the Van Herick procedure and an Artificial Intelligence approach for estimating the angle aperture. In this work, we propose a deterministic and quantitative vision-based algorithm for the evaluation of the Anterior Chamber Angle aperture. The proposed algorithm allows the estimation of the Van Herick grade from 1 to 4 by computing the ratio value between the Anterior Chamber Depth and the Corneal Thickness. The algorithm is evaluated on an image dataset acquired from patients and classified by expert ophthalmologists. The results show an agreement between clinical classification and the algorithm estimation higher than 65 %, which reaches 100 % for grades 4. Moreover, the algorithm provides the numeric value of the ratio between Anterior Chamber Depth and Corneal Thickness, which can be used as new quantitative information about the angle closure.