Image Edit Based On Re-texturing With Interactive Extracting

Image edit is a very important research direction in image related field and it is a process that outputs the satisfied processing result according to the editing require of the user. With the increase of image editing requires, research of different and novel image editing methods become the key point in the image and graphic field. The work done in this paper derive from the practical require of certain cooperated image and game designing company which hope to create new picture or photo via changing the texture appearance of the object and to simplify the process of make new pictures. At present, editing software generally allows images to be cut/inpainting, zoom/rotate, color-check. But these functions can not satisfy require of the texture editing. Therefore, an interactive re-texturing editing is designed and implement and this paper pay more attention to interactive extracting and re-texturing research.In the field of extracting techniques, lots of researches and comparisons are done among various relative techniques of extracting and we focus on the interactive extracting method. For getting a good alpha matte, the extracting function is extended and constrains are imported into the cost function for get a good alpha matte under user's guiding. The algorithm only needs simple user interaction while remains the extracting accuracy, and it is convenient and quick to get a good extracting result.In the field of extracting techniques, two re-texturing methods are analysed and compared. Re-texturing based on texture replace is chosen to design and implement the interactive re-texturing editing. The re-texturing editing method combining extracting object interactively only needs several scribbles which indicate the foreground object and the background. Then an exact alpha matte can be obtained and the object can be re-textured well. In re-textured images or photos, sometimes the whole image has some color harmonization problem due to the change of the texture. For solving this problem, colorization method based on original image is presented and it can make real and natural re-texturing results.According to the results of experiments, interactive re-texturing editing presented in this paper meets the project requires well. Additionally, the improvements of related technology make the whole editing process flexible and meet user's need of edit and creation. Hence the method presented in this paper can have wide application throughout photograph editing, art design and visual effects.