Face Recognition Based on Phase Only Correlation (POC)

Abstract

Face recognition is a technology that is widely used in the field of security and biometrics. Every human face has unique features such as the shape of the jaw and the contours of the face. In this study, face detection was carried out using the Phase Only Correlation (POC) method. POC is a face detection method based on the highest correlation value from the calculation of the phase and magnitude of an image. The data used in this study is in the form of facial images in RGB format. Stages of the study were divided into two, namely preprocessing and processing. The pre-processing stage includes the Region of Interest (ROI), image segmentation, grayscale transformation, and Discrete Fourier Transform (DFT). The processing phase includes Cross Spectrum, Inverse Cross Spectrum values, and correlation calculation. There are three testing schemes carried out in this study, namely testing of variations in lighting levels, expressions, and facial positions. From the results of the study, different face positions produce incorrect detection results because changes in the position of the image being tested make changes in the frequency distribution of the phase and magnitude values.