Comparative Study: Face Recognition via the Correlation Filter Technique

Face recognition attracts much attention in various applications due to its non-intrusive nature and the widespread availability of digital cameras. Recently, the benefits of using spatial frequency domain representations for face recognition have drawn great interests from the computer vision and pattern recognition community. In this paper, we present a comparative study by using the correlation filter (CF) technique in the application of face recognition. We overview some representative correlation filters (CFs) proposed recently and analyze their respective pros and cons. Experiments using different types of CFs with different training parameters are conducted on public face databases to investigate the overall performance of the CF-based face recognition methods. The observations based on these experiments are expected to provide widely applicable guidelines for designing the face recognition systems via the CF technique.