Implementation and testing of local binary pattern histogram as a method of face recognition

Abstract

The popularity of the cameras in smart electronic devices led the industries to use them more efficiently. Facial recognition research is one of the favorite topics among practitioners and researchers and is a key to the future of smart technologies. This study is an attempt to indicate the effectiveness of existing facial recognition algorithms using OpenCV library and C# programming language. This thesis aims to investigate several facial recognition algorithms and make comparisons in respect of their accuracy. We will use Viola-Jones Face Detection algorithm for detecting the face, and the Eigenfaces, Fisherfaces, Local Binary Pattern Histogram algorithms for recognizing the face. The thesis covers the complete process of face recognition, including face detection, preprocessing of images, the comparison of the algorithms mentioned above and the real-time application of Local binary pattern histogram. We will discuss the concept of each algorithm, and comparative analysis will reveal the most accurate one. The development of the test cases will indicate that among the compared facial recognition algorithms the Local Binary Pattern Algorithm has the highest accuracy to identify faces.