Glaucoma detectıon and classıfıcatıon by usıng hysteresıs thresholdıng based IAOAWTO and ICNNBTL classıfıer

Nowadays, a lot of patients undergo eye screening per day, therefore ophthalmologists face a lot of challenges during the screening of glaucoma. Also, manual screening leads to errors and is more time-consuming, the patients have to wait for much time in the clinic. Hence an automated system is essential to help the ophthalmologist as a secondary opinion in retinal screening. In this research work, the initial step is image acquisition; here the input images are collected from public ORIGA and in-house Clinical Fundus Images. The next step image enhancement is performed using a Bilateral with Unsharp Filter, which can eradicate the noise discerned in the input images. In the subsequent step, the Hysteresis thresholding with the morphological post-segmentation process is implemented on an enhanced image sequence for Retina Blood vessel extraction. Then the segmented output is fed to the Improved Archimedes Optimization Algorithm with Transfer Operator (IAOAWTO). Ultimately, the Fundus images are classified by utilizing an improved convolutional neural network-based transfer learning (ICNNBTL) classifier. The accuracy is 96.6% and 5 FP is predicted. The results show the improvement of diagnosis in the proposed method compared to other methods.