The research conducted is testing the accuracy of the level of similarity of the management of STMIK Pelita Nusantara. The facial images tested were 17 images and 136 tests in each method (Laplacian of Gaussian (LoG), Canny, and the combination of LoG + Canny). Tests were carried out using Matlab R2017b. From the test results, the researchers concluded that the accuracy of the highest level of similarity is the Laplacian of Gaussian method, namely images 12 and 17 with a percentage of 99.85%, then the Canny method, namely images 4 and 7 with a percentage of 99.53% and the lowest is the combination of the two methods. (LoG + Canny) namely images 6 and 13 with a percentage of 98.14%. And the highest average accuracy of the similarity window is the Laplacian of Gaussian method with a percentage of 49.91%, then the Canny method with a percentage of 38.19% and the lowest is the combination of the two methods (LoG + Canny) with a percentage of 37.81%.
Copyrights © 2021