Face Recognition Accuracy Improving Using Gray Level Co-occurrence Matrix Selection Feature Algorithm

Abstract

Information security is becoming increasingly important as technology develops, but it also faces many challenges. Face recognition is one of the technologies that can be used to face these challenges. The facial recognition system has been studied extensively, but its accuracy has not been optimized. Therefore, further studies are needed to increase accuracy. The feature extraction process can be used to improve accuracy. This method is used as a reference to extract important information in an image and distinguish one image from another. Several feature extraction methods are available for face recognition, one of which is the Gray Level Co-occurrence Matrix (GLCM) method. This research goal to improve the accuracy of facial recognition by trying a number of features in GLCM. Face recognition uses Backpropagation because this research only proves the effect of GLCM features, not to find maximum accuracy. Based on the trials that have been carried out, the best test accuracy results without GLCM feature extraction are 73%. While the test using extraction with 4 GLCM features is 89% when the GLCM neighbor distance is 1 pixel, so there is an increase in accuracy of 16%.