A study of the eigenface approach for face recognition

Appearance-based approaches in face recognition, specifically the Eigenface approach, were one of first successful demonstrations of machine recognition of faces [1]. These methods, such as those proposed in [2, 3], proved to be effective in experiments with large databases. Further development of holistic methods of face recognition and their theoretical background, such as those proposed in [4-9], were focused towards recognizing faces from images with changes caused by illumination effects and pose variations. Although much effort has been made towards this goal, current algorithms are still far away from the capability of the human perception system [1]. This report shall give a detailed description of the fundamentals of appearance-based holistic methods for face recognition, specifically the Eigenface approach [3], as well as our experimental results on the Yale Face Database