| Summary: | Glaucoma is the second leading causes of blindness beside of cataracts. In the glaucoma examination, determining the disease is very difficult since the symptoms are difficult to find. The ophthalmologists usually use the disc damage likelihood score (DDLS) in examining the optic nerve damage which is very complicated. Glaucoma can also be detected by looking at damage of retinal nerve fiber layer (RNFL) which is characterized by the loss of nerve fiber around the optic nerve head (ONH). However, manual examination became less effective and less efficient since it required specific skills and needed more time to finish the examination. This research aimed to develop a scheme for detecting glaucoma disease based on retinal nerve fiber layer features. First order statistic feature was used to recognize the characteristic of RNFL. These features were mean, standard deviation, skewness, energy, entropy, smoothness, min and max. The info gain feature selection was used to select features with the most dominant influence. Then, the selected feature was used to classify the image into a normal class or glaucoma. Retinal image classification results between normal class and glaucoma using radial basis function network (RBFN) classifier obtained accuracy of 90.00, sensitivity of 93.33, and specificity of 86.67. These results indicated that the proposed method could assist the ophthalmologist in early detection of glaucoma disease on retinal fundus image based on retinal nerve fiber layer features. © 2021 IEEE.
|
|---|