Glaucoma is a domineering and irretrievable neurodegenerative eye disease produced by the optical nerve head owed to extended intra-ocular stress inside the eye. Recognition of glaucoma is an essential job for ophthalmologists. In this paper, we propose a methodology to classify fundus images into normal and glaucoma categories. The proposed approach makes use of image denoising of digital fundus images by utilizing a non-Gaussian bivariate probability distribution function to model the statistics of wavelet coefficients of glaucoma images. The traditional image features were extracted followed by the popular feature selection algorithm. The selected features are then fed to the least square support vector machine classifier employing various kernel functions. The comparison result shows that the proposed approach offers maximum classification accuracy of nearly 91.22% over the existing best approaches.

青光眼是一种霸道且无法治愈的神经退行性眼病，由视神经乳头由于眼内长期眼内压力而产生。青光眼的识别是眼科医师必不可少的工作。在本文中，我们提出了一种将眼底图像分类为正常和青光眼类别的方法。所提出的方法通过利用非高斯二元概率分布函数对青光眼图像的小波系数的统计进行建模来利用数字眼底图像的图像去噪。传统的图像特征被提取，然后是流行的特征选择算法。然后将所选特征馈送到采用各种核函数的最小二乘支持向量机分类器。