UJI KETAHANAN IMAGE WATERMARKING DARI METODE CRT DENGAN DETEKSI TEPI CANNY UNTUK CITRA RABBANY

Abstract

One technique to obtain validity in protecting a brand is with a watermark, one of the watermark algorithms to provide the validity of a brand is the Chinese Remainder Theorem (CRT) algorithm. In previous studies using the CRT method with canny edge detection has the ability that the watermark produced is almost similar to the original image and the results are quite good, the CRT method in canny edge detection can significantly improve the quality of the watermark image with an average of 0.99942. Whereas when given an image processing attack that is noise 'salt and pepper' and jpeg compression can maintain robustness with an average compression of 0.66187 and noise 0.940243. However, this method is not imperceptibility and robustness against Gaussian blur cropping, rotate 30 °, rotate 45 °, rotate 90 ° and rotate 180 ° where the average NC value away from 1. From the results of the research proposed the method tested has not been imperceptibility and robustness against Median Filter attacks, Contrast times 1.5, Contrast times 1.8, Scaling 359x359 pixels and Scaling 256x256 pixels where the average NC value is still away from 1. But in Median Filter and Scaling attacks 256x256 pixels indicate average results of extracted attacks can still be relatively identified with the naked eye even though the NC value generated does not approach 1.