A Machine Learning based Approach for Segmenting Retinal Nerve Images using Artificial Neural Networks

摘要

Artificial Intelligence (AI) based Machine Learning(ML) is gaining more attention from researchers. Inophthalmology, ML has been applied to fundus photographs,achieving robust classification performance in the detection ofdiseases such as diabetic retinopathy, retinopathy of prematurity,etc. The detection and extraction of blood vessels in the retina isan essential part of various diagnosing problems associated witheyes, such as diabetic retinopathy. This paper proposes a novelmachine learning approach to segment the retinal blood vesselsfrom eye fundus images using a combination of color features,texture features, and Back Propagation Neural Networks(BPNN). The proposed method comprises of two steps, namelythe color texture feature extraction and training the BPNN to getthe segmented retinal nerves. Magenta color and correlation-texture features are given as input to the BPNN. The system wastrained and tested in retinal fundus images taken from twodistinct databases. The average sensitivity, specificity, andaccuracy obtained for the segmentation of retinal blood vesselswere 0.470%, 0.914%, and 0.903% respectively. Results obtainedreveal that the proposed methodology is excellent in automatedsegmentation retinal nerves. The proposed segmentationmethodology was able to obtain comparable accuracy with othermethods.