Research Of Image Editing Method Based On Digital Photomontage

Digital photomontage is an image editing technique that starts with a set of different picture taken from the same scene, and goes through composite and fusion process to create a single photograph that better conveys the photographer's perception of the scene. As a new image based rendering technique, digital photomontage has been studied widely and deeply in recent years. It has an extensive research foreground. However, as a newly appeared technique, the existing digital photomontage wasn't perfect enough at present. Therefore, it's necessary to do further research on it.In this paper, we introduce the research and applications on the image editing method of digital photomontage. The primary technical challenges of photomontage are 1) to choose good seams between parts of the various images so that they can be joined with as few visible artifacts as possible;and 2) to reduce any remaining artifacts through a process that fuses the image regions. We use the max-flow/min-cut algorithm based on graph-cut based energy minimization theory to solve the first problem. We also use gradient-domain fusion (based on Poisson equation) to reduce or remove any visible artifacts that might remain after the image seams are joined.We introduced two kinds of image editing methods: interactive digital photomontage and non-interactive digital photomontage. We first explain the different processing and techniques of the two methods. Then we introduce the implementation detail of graph-cut based energy minimization method and the Poisson image editing method for these two kinds of situations.We developed an image editing system of digital photomontage, including interactive and non-interactive. For the non-interactive application functions, we mainly realize extended depth of field, stroboscopic visualization of movement, background reconstruction and clean plate. For interactive method, we realize selective composite and facial synthesis function. In facial synthesis module, we provide the function of creating portrait database. Through simple user selection, our system can first choose the best face parts, and then composite and fuse them automatically. Users can also search and update the database. The experiment results show that this system can be applied to many applications.