IMPLEMENTATION OF CONVOLUTION MATRIX AND PSNR ON BRAIN TUMOR IMAGE

Abstract

Brain Tumor is a deadly disease that affects both men and women in this life. Detection of brain tumors can be using CT scan (Computed Tomography scan) and Magnetic Resonance Imaging (MRI). In this study took 5 brain images from CT scans and processed into 25 images. The algorithm used is 3x3 kernel Matrix Convolution and 5x5 kernel for image sharpening and blurring, after which the values of MSE, RMSE and PSNR will be obtained. The best result of convolution 3 brain tumor image obtained from the image input image sharpening kernel 3x3 is the first, the image input 5 with the value of MSE 1308.597034, RMSE 36.174536 and PSNR 16.962744 dB. Second, with the input image 3 with the value of MSE 1316.962532, RMSE 36.289978 and PSNR 16.935069 dB. Third, the input image 1 with the value of MSE 1325.889702, RMSE 36.412768 and PSNR 16.905730 dB. The best result of convolution 3 brain tumor image obtained from the input image blurring kernel 5x5 is the first, the input image 5 with the value of MSE 2235.845032, RMSE 47.284723 and PSNR 14.636387 dB. Second, with the input image 4 with the value of MSE 2291.064041, RMSE 47.865061 and PSNR 14.530431 dB. Third, the input image 3 with the value of MSE 2292.273529, RMSE 47.877693 and PSNR 14.528139 dB.