Design of Face Recognition Detection Using the Local Binary Pattern (LBP) Method

Abstract

Face recognition on an Android smartphone is a personal identification system that uses the physical characteristics of a person to automatically detect their identity through faces. In this face recognition detection technology will identify human faces through a digital image by matching or comparing with faces that have previously been stored and trained in data storage.

The main problem faced in designing face recognition detection is how android devices can detect and recognize face objects. The proposed design method for Android-based face recognition detection is the local binary pattern (LBP) method, which in principle uses this method to get the characteristics of each face image by representing a gray image into binary numbers. The method developed is by modifying the heuristic function of the Ant-Miner algorithm.

From the test results obtained that the performance of facial recognition detection using the LBP method on a person's face image has a good level of accuracy with an average of 96.7%, with 5 trials and 30 trials. In addition, this face recognition detection can also be used as a test framework for Android-based face recognition detection.