From coarse to fine skin and face detection

Abstract

A method for fine skin and face detection is described that starts from a coarse color segmentation. Some regions represent parts of human skin and are selected by minimizing an error between the color distribution of each region and the output of a compression decompression neural network, which learns skin color distribution for several populations of different ethnicity. This ANN is used to find a collection of skin regions, which is used in a second learning step to provide parameters for a Gaussian mixture model. A finer classification is performed using a Bayesian framework and makes the skin and face detection invariant to scale and lighting conditions. Finally, a face shape based model is used to decide whether a skin region is a face or not.