Automatic face detection based on chrominance components analysis
Marian Beniak, Jarmila Pavlovičová, Miloš Oravec · 2008
This paper describes our approach to face localization task. In the first part we discuss the way to find a suitable color model for automatic face segmentation and in next part, there is a practical application of the model in an algorithm which major use is in the videophone applications according to his simplicity and technical compatibility with videophone technology (YCbCr color space). This approach leads us through pixel classification based on chrominance components values of pixels, than we perform morphological operations, in next step we improve the results with help of chrominance parameter of pixels. Using morphological operations we enhance the accuracy of the results and in last step algorithm gives us a map of the region in which face is located.