Performance comparison of some face recognition algorithms on multi-covariate facial databases

Facial recognition based systems are ubiquitous. From domestic to industrial usage, these automated methods have become an essential part of this digital world. The problem of face verification has been exhaustively researched, thereby producing some robust and accurate face recognition algorithms. Our present work investigates and relatively analyzes the effectiveness of eight such algorithms which have been widely studied and implemented in the research community. The performances of these techniques have been evaluated on six real-life facial databases, thereby providing baseline results for each one. Thus our comparative study not only analyzes the usefulness of each method but also would (hopefully) provide guidelines in designing application specific facial recognition techniques in the near future.