A Colorization Method Of Gray Images Based On Color Mixing

Images are visual entities, which can be obtained by observing the world through various observation systems and can be directly or indirectly raised in human eyes. Visual information holds most in all kinds of information which get from the outside (about 80%). Color and gray are key factors in determining the presentation of an image, and the former plays an important role in human visual system. For some gray image, such as:infrared photos, gray pictures, and magnetic resonance imaging, To better the visual effect by turning the gray images into colored images, the colorization processing of image is to be used.The process of coloring images is an ill-posed problem, so we need add specific conditions to the process, through various programming methods making sure this transformation is unique. According to the methods of processing, there are two kinds of colorization:pseudo colorization and false colorization. pseudo colorization refers to the objects which have not colors, by establishing the corresponding relationship of grayscale to the color, changing the gray images into color images, false color objects have color, but in the process of imaging, their color missing, so we should simulate the real scenes using colors.First, the paper describes various colorization algorithms for recently years, then given a detailed introduction about Yatziv algorithm which based on intrinsic distance. On the foundation of this, author use the colorization method based on the unevenness, and reduces computation cost by decreasing and increasing the resolution of color space, means retaining the information of brightness and compressing the color information. Through all kinds of tests, we found that the method can get a good effect. For the feature of the static background video sequence, author designed a new method of colorizing video sequences. Colorize the background and moving object that are accessed, and then colorize the next frames, according to the correspondence between pixels. Combined with the colorized background, we can get a colorized video sequence.