Face detection has known as the most essential and primary process in the field of many different application like face recognition, biometric authentication, database contextual retrieval, more sophisticated human-computer interaction and so on. In the last decade the attempts for facial detection have faced to color images. Several different researches and challenges have been done to classify the image regions into two groups of skin and non-skin pixels. In this work in spite of some former researches, considering the fact of importance of luminance as well as chrominance elements, the complete color information has been used for this purpose. It is also tried to cover all different kind of skin colors from completely dark (black) skins to yellow and white ones. Application of LVQ neural networks has enabled the system to overcome the complexity of distribution of samples. The designed model shows more accurate results in almost all cases and also the flexibility of LVQ classifier makes it possible to apply in different color spaces
展开▼
