This paper proposes an improved version of ourpreviously introduced face detection system based onskin color segmentation and neural networks. Thenew system, using a support vector machine (SVM)based method for learning and verification, consistsof several stages. First, the system searches for theregions where faces might exist by using skin colorinformation and forms a so-called skin map. Afterperforming noise removal and some morphologicaloperations on the skin map, it utilizes the aspect ratioof a face to find out possible face blocks, and theneye detection is carried out within each possible faceblock. If an eye pair is detected in a possible faceblock, a region is cropped according to the locationof the two eyes, which is called a face candidate;otherwise it is regarded as a non-face block. Finally,each of the face candidates is verified by a supportvector machine. Experimental results reflects that thenew version improves the verification accuracy of thepreviously proposed system.
Relation:
第十三屆人工智慧與應用研討會論文集=The 13th conference on artificial intelligence and applications, pp.158-164