Invariant Gabor Features for Face Evidence Extraction
Joni‐Kristian Kämäräinen, Ville Kyrki, Heikki A. Kalviainen, Miroslav Hamouz, Josef Kittler · 2002
Invariant feature extraction is one of the most difficult problems in machine vision. Human face detection and recognition have recently become an important application area of computer vision. In this paper, a novel illumination, translation, rotation, and scale invariant feature extraction method based on Gabor filtering is introduced. The proposed method is successfully applied to invariant detection of facial features. In addition to face detection, the proposed theories and methods can be applied to a wide variety of object detection problems. 1