An Application of Gabor Filters for Texture Classification

In the contribution we present the evaluation of texture segmentation efficiency for Gabor filter bank and we bring an application to face feature detection. Gabor filters are well known as a facility for texture because of their orientation and spatial frequency character. We use several statistics to extract more information from Gabor-filtered images. We applied variance, energy and percentiles as a statistics useful for human hair detection and classification. We found this face feature useful for gender identification, but also as a supplementary feature for face detection.