SEGMENTASI PEMBULUH DARAH RETINA MENGGUNAKAN MORPHOLOGY FILTER DAN ISODATA
Daftar Isi:
- The retina is one of the locations in the human body where blood vessels play an important role in the diagnosis of eye diseases. One technique for diagnosing diseases through blood vessels is retinal image segmentation. The ISODATA thresholding method is a method for segmentation that uses iteration that can calculate the average level of gray, and can automatically find the threshold value of an image. this study also uses morphological operations for optical disk removal and the dataset used are 20 DRIVE test images with ground-truth. The results of this study are seen from the performance of the system designed based on the parameters of accuracy, sensitivity, and specificity. The accuracy obtained in this study was 94,956%, sensitivity was 82,471%, and specificity was 96,560%. For the results of testing parameters based on the level of similarity of the image using MSE, RMSE, and PSNR parameters. MSE results obtained 4,590, RMSE of 2,124 and PSNR of 41,700 dB. For image signal quality, testing using SNR parameters by comparing the noise image is CLAHE image, and the filtered image using 2d median filter, the average value of each noise image is 24.01 dB, and filtering image 24.21 dB.