Estimation of Clinical Parameters for Glaucoma Detection Using Image Processing and Deep Learning Techniques from the Biomedical Human Eye Images

The technique for distinguishing the essential glaucoma is being exhibited in this segment utilizing a changed neural system with the back spread calculation ideas. The morphological administrators ideas are being utilized for the handling of the cup and disc, in the sense, the cup and the disc zones are discovered from which the proportion is registered, from where the illness can be recognition seeing the proportion. Utilizing an updated highlight extraction process, the highlights of the disc portion and the cup can be discovered, which could be utilized in the idea of cup to plate apportion (cup to disc ratio) recognition, regardless of whether the given test image is glaucomatic or not. Classification is also done using artificial neural networks (ANN) along with the training process. Simulations have been done in the Matlab environment by developing. m codes. Once the developed program is run, the simulation results are observed. From the simulation results, it can be observed that using the computed values of the cup to disc ratio, the severity of the disease can be computed (whether it is normal or moderate or severe glaucoma). Databases have been taken from the hospitals and on-line standard websites. The quantitative results are tabulated in a table. Finally, it was found that after taking a survey of 50 images, 16 were affected with normal, 16 were affected with moderate and the rest were affected with severe case of glaucoma. The Matlab results of the simulation done using coding demonstrated the profoundness with respect to the strategy proposed or developed in this exploration write-up using the concept of neural networks.