Study On Image Segmentation And Composition

In an image, we are often interested in some part, also called object, which has specific statistics. To analysis and understand the object, it is necessary to segment it from its image. Because of the importance of imge segmentation in practice, it is an important research field in computer graphics and there are many methods for it. In this paper we propose a new image segemantation method. The method first separates these pixels which obviously belong to object or background, then gets the cutout whose corresponding value of the new energy function is the minimum. After segmenting the object wanted, one of the applications is fusing it into another background to get a new image, like science film and cartoon. Because the scenes in these images don't exist actually or are too costly to construct, we cannot get these images by shooting. Fortunately we can obtain these images through imge composition which merges several parts from several images into an imge, however if the composite image is valuable, it must appear integrative and cannot have visible montage marks. This paper puts forward to a gradient simulating method. The gradient simulating method ensures the compliance of illumination in the composite image meanwhile preserves the color characteristics of the object. The experiments showed that the new composition method has extended applications and the composite image has a better lighting than those produced by other methods.