The Research And Improving On A Matting Method Based On Iterative Imaging Segmentation

Digital image matting is to extract the foreground objects with arbitrary shapes from an image. It is an essential technique to make special effects in film and video production and has enormous potential commercial applications. The existing image matting techniques can't satisfy the increasing needs for speed and effects of matting simultaneously. This paper proposes optimal and developed algorithms for Grab Cut algorithms。First, research the Modal and formula of the Grab Cut algorithm.With not influencing the result of the original algorithm, improved a part of the Modal and formula of the original algorithm.And we quote the clustering algorithm of Color Quantization of Images to deal with the initialization of GMM. Comparing to the other algorithm, we will get a better result.Second, In the procedure of iterated computing, this paper proposes the re-initializing of GMM and the mix initializing of GMM in the procedure of iterated computing to deal with the thing that the component of GMM disappear in the procedure of iterated computing. Comparing to the original algorithm, this mode will improve the speed of convergence and the last result of segmentation.Third,in order to deal with the problem that the minimization algorithm presented in Grab Cut paper which only deal a pixel and the pixels around it in each computing,only guarantees convergence to a local minimum, we present a method to restrain the global minimum by using Inter-differ pair of dot.