Region-Growing-Based Poisson Matting

The matting problem is to extract a foreground element of arbitrary shape from a background image by estimating a color and opacity for the foreground element at each pixel. It has been widely used in Image Processing, film and video production, and Virtual Reality. In natural image matting, researcher can get alpha value through statistic, resolving differential matting equation and other methods, after the researcher drawing the foreground and background contours. And then he can get the weight of the foreground color in the unknown region and drawing it out. Better result or fast processing is the main purpose of this paper.The basic theories and knowledge of digital image matting is introduced in the paper at first. Then it introduces several kinds of matting methods in common use present. A Region —Growing —Based Poisson matting is proposed based on Poisson matting, and natural image matting is realized using this method.Region Growing is a basic Image Segmentation method. First, it takes a given point as the seed. Then the surrounding pixel is merged into the area that the seed is contained. In Human Vision System, the optical signal is separated to brightness and chroma information, and they are processing differently. It is an important quality of Human Vision System. The maximum vector of the RGB color in the RGB color space determines the sense of Human Vision System to the color. The Human Vision System reflects the perceptual character of human eyes.Combine the Human Vision System theory with Region Growing, color images is segmented more efficient. Using the method to Poisson matting, less processing time and better processing result is gained.