Human retinal optic disc detection with grasshopper optimization algorithm

A growing number of qualified ophthalmologists are promoting the need to use computer-based retinal eye processing image recognition technologies. There are differ- ent methods and algorithms in retinal images for detecting optic discs. Much attention has been paid in recent years using intelligent algorithms. In this paper, in the human retinal images, we used the Grasshopper optimization algorithm to implement a new automated method for detecting an optic disc. The clever algorithm is influenced by the social nature of the grasshopper, the intelligent Grasshopper algorithm. Include this algorithm; the population contains the grasshoppers, each of which has a common luminance or exercise score. In this method, two-by-two insects are compared, so it could be shown that less attractive insects shift towards more attractive insects. Finally, one of the most attractive insects is selected, and this insect gives an optimum solution to the problem. Here, we used the light intensity of the retinal pixels instead of grasshopper illuminations. Accord- ing to local variations, the effect of these insects also indicates different light intensity values in images. Since the brightest area “represents the optic disc in retinal images, all insects travel to the brightest area, which leads to the determined position for an optic disc in the image. The performance was evaluated on 210 images, reflecting three Open to the public and sequentially distributed datasets DIARETDB1 89 images, STARE 81 images, and DRIVE 40 images. The results of the proposed algorithm implementation give a 99.51% accuracy rate in the DiaRetDB1 dataset, 99.67% in the STARE dataset, and 99.62% in the DRIVE dataset. The results of the implementation show the strong capacity and accuracy of the proposed algorithm for detecting the optic disc from retinal images. Also, the recorded time required for (OD) detection in these images is180.14 s for the DiaRetDB1, 65.13s for STARE, and 80.64s for DRIVE, respectively. These are average
values for the times.