Gradient Domain Based Processing Method for Image Synthesis

Abstract

Digital image synthesis technology is an important technology in the field of computer graphics. It is widely used in various scenes such as digital image editing, film and television production, graphic design and so on. Gradient domain is an important concept in digital images. It can reflect the degree of variation of pixel values in each location of a digital image in its adjacent areas, and has a good perception effect on the edges of objects in digital images. Gradient-domain image synthesis method is an effective technique in the field of digital image synthesis. Its goal is to select a specific region from the source image and fuse it into the background image so that the images which before and after the fusion have no fusion traces. Vivid and natural. In this paper, the characteristics of digital image synthesis method based on gradient domain are analyzed. The Poisson clonal algorithm is used for image fusion. The characteristics of the core equations involved in Poisson clonal algorithm are studied. The equations are solved by combining the characteristics of coefficient matrix. The process is optimized and the solution of the linear equation is replaced by a Gaussian elimination method with a conjugate gradient method, which enables the digital image fusion process to be completed within a user-acceptable speed range.