An Enhanced LBPH Approach to Ambient-Light-Affected Face Recognition Data in Sensor Network

Although combining a high-resolution camera with wireless sensing network is effective for interpreting different signals image presentation on the identification of face recognition, its accuracy still severely restricted. Removing unfavorable impact ambient light remains one most difficult challenges in facial recognition. Therefore, it important to find an algorithm that can capture major features object when there are changes. In this study, recognition used as example analyze differences between Local Binary Patterns Histograms (LBPH) and OpenFace deep learning neural algorithms compare error rates environmental lighting. According prediction results 13 images based grouping statistics, rate LBPH higher than scenes changes When azimuth angle source more +/−25° elevation +000°, low. +25° −25° higher. Through experimental design, show that, concerning uncertainty illumination angles lighting source, has