The Study, Based On Face Detection And Feature Verification Algorithm

Face recognition is a complex task that has received a great deal of attention in recent years.lt has many applications,e.g.person identification,criminal investigation and security surveillance etc.Face detection as the primary preparation for face recognition also plays an important role.Human face detection is to locate the faces and determine their sizes in an image.Through introducing the basic of face detection description,this article mainly has a through research on the algorithms of face detection technology.Then two new face detection methods are presented.And as facial feature has an extraordinary significance in face detection,feature detection is proposed at the last.(1)A face detection algorithm based on color skin and similarity is determined.A new illumination compensation system is used,by comparing the various color space,the YCbCr color space is choosed and improved,then the similarity of color skin is calculated,finally the face region is made up in the image.(2)A new method based on the elliptical model to locate precisely the contour of the face is presented.through a series of preliminary processing of image,the Canny operator is selected and improved.And a new way based on the edge-linking is advanced to removal the short edge and pseudo-edge.Thus the edge image with good result is obtained.According to the oval template,the face is detected.(3)Because eyes are considered the most salient and stable among all other facial features,a novel and robust eye location algorithm is proposed,which combine the radial symmetry transform and gray-level integral integral projection to locate the eyes accurately.All the experiment results are very good in the laboratory environment.