Effects of Preprocessing Eye Fundus Images on Appearance Based Glaucoma Classification

Abstract

Abstract. Early detection of glaucoma is essential for preventing one of the most common causes of blindness. Our research is focused on a novel automated classification system based on image features from fundus photographs which does not depend on structure segmentation or prior expert knowledge. Our new data driven approach that needs no manual assistance achieves an accuracy of detecting glaucomatous retina fundus images compareable to human experts. In this paper, we study image preprocessing methods to provide better input for more reliable automated glaucoma detection. We reduce disease independent variations without removing information that discriminates between images of healthy and glaucomatous eyes. In particular, nonuniform illumination is corrected, blood vessels are inpainted and the region of interest is normalized before feature extraction and subsequent classification. The effect of these steps was evaluated using principal component analysis for dimension reduction and support vector machine as classifier.

Automatic extraction of retinal features from colour retinal images for glaucoma diagnosis: A review

“…In non-segmentation based methods, Bock has major contribution [95,96,92]. Initially they used Pixel Intensity Values on which they applied Principal Component Analysis (PCA) [97] for dimensionality reduction.…”

Classifying Glaucoma with Image-Based Features from Fundus Photographs

“…Our previous studies [15] showed that appearance-based approaches perform better on images with less disease independent variations. For this reason, a normalization of the three major variations is applied.…”

“…We considered principal component analysis (PCA) as an unsupervised and linear discriminant analysis (LDA) as a supervised method to reduce dimensionality. As evaluated in [15], thirty principal components capture at least 95 % of data variation.…”

Fundus image analysis for automatic screening of ophthalmic pathologies.

“…In (Rathinam and Selvarajan, 2013) basic pre-processing techniques such as contrast adjustment, adaptive histogram equalisation, median filtering and average filtering are applied before glaucoma detection. In most of the state-of-the art works (Aliaa et al, 2007;Marrugo and Millán, 2011;Meier et al, 2007;Rathinam and Selvarajan, 2013), the non-uniform illumination is the main correction explored. In (Aliaa et al, 2007) a colour normalisation based on histogram equalisation and shape matching is applied on each RGB component separately with the aim of detecting the retinal anatomy.…”

“…Different authors (Giachetti et al, 2011;Meier et al, 2007;Morales et al, 2013) have inpainted blood vessels by using diffusion-based algorithms but only as a middle step of their purpose, in other words, without any kind of evaluation of the inpainting method used. For example, Morales et al (2013) uses inpainting techniques in order to segment the optic disk in fundus images.…”