Theory And Algorithm Research On Image Morphing

Image morphing technology is a new branch of image processing in recent years, which studies space mapping of pixels and pixels in images. It can come into being special effects using image morphing, which has been widely used in the fields of facial detection, facial image synthesis, biomedicine image, film, television and advertising media, etc. With the development of computer graphics, the method of image morphing has been developed increasingly and has been widely used in many fields. Image morphing can transform one image to another in a natural, fluent, vivid way. In the process of image morphing, starting image is warped and fade-out gradually, while ending image is fade-in. In the whole process, in-between images take on the characters of both starting image and ending image. The thesis taken a deep research on the edge detection method of digital image, image morphing based on feature-lines and image morphing algorithm based on optimal mass transport problem etc, and experiment result is given.In the edge detection of image, Wavlet transform method is used to get polar value by multi-scale decomposition to realize multi-scale edge detection.In the image morphing based on feature-lines, Hough transform and feature-lines is used to realize the morphing. And the method of getting median of the lines is also used to get the ultimate morphing image by color superposition.In the algorithm of auto image morphing based on optimal mass transport problem, we applied a modified Monge-Kantorovich flow to the problem of image morphing by adding a comparison term to the optimal mass transport energy functional. Using cross-dissolving method, we got the in-between images sequence, We had a satisfied result ultimately.