Research On Algorithm Of Automatic Face Image Beautification

With the development of science and technology, more and more digital products have a good performance in displaying the objects. Take smart phones, video equipment,digital camera for examples they can all display well enough to show the objects well.People often use the digital camera and smart phone to take pictures. Since most people have deficiencies more or less. People never give up the pursuit of beauty, we need a system which could optimize our faces. Thus, the software such as Ever Fine and PS comes out, but the user must know the knowledge of the software and the operation is also complicated. So we need a automated system which could optimize our faces undistortedly.This paper realizes the algorithm of the face deformation based on face feature points detection.It also accomplish the face deformation system. Firstly, we should detect out the face region based on skin color model.Then, we should point out the position of the cornr of the eyes and the corner of the mouth, nostrils, etc by the way of PCA validation and color segmentation. After the compare of several morphing algorithms, we choose thin plate splines deformation algorithm to realize the deformation system. Because this algorithm often make the image twisted in the unexpected area, we need increase the control points to make the image stable. It will increase the calculation and affect the system's real-time. This paper puts forward the thought that we deform the face features only in a rectangular area,which make the system real-time and effective.This paper realizes the algorithm of the face deformation in the pictures of human face under the environment of VC++. The result shows that this system could beatify the image of human faces quickly, and this system has the value of practical applications.