Research Of Digital Image Inpainting Technology

As an important branch in the field of image processing, image restoration involves pattern recognition, mathematical statistics, information theory, computer graphics, machine vision and many other fields. As the focus topic, it has been widely used in military, medicine, industry, photography, archeology, games, and special effects production in films and other fields. The purpose of image inpainting is to repair and restore the missing information in the image according to the image'characteristics and the information around the damaged area, recover the lost information, make the image more clear and natural, and maintain the overall visual effect of image.In this paper, we introduced the research background and value of image inpainting technology, gave the mathematical description and summarized the research status in different stages. Then we studied the PDE-based image inpainting techniques, discussed several classic models in detail, such as BSCB model, TV model and CDD model. Besides that, by giving two threshold parameters to judgment the weighted value in the iterative formula, we presented an improved algorithm based on the TV model. Simulation results shown that this algorithm can not only improve the quality of image restoration, but also effectively reduce the unnecessary duplication of iteration, and consequently improved the efficiency of the repairing algorithm.In this paper, we discussed image restoration based on texture synthesis techniques, and proposed a gradual texture synthesis algorithm based on the traditional exemplar-based method. By analyzing the image's structure and texture characteristics, we firstly set the priority for repairing based on the confidence function and the structural function. Secondly, we calculate the priority of each points on the edge, thirdly determine the repairing priority order according to the priority number of each points, the point with high priority need to repaired in advance. Finally select the best match according to the certain rules or algorithm, and add the information in best matching block to the matched block directly, until complete the restoration of entire image.By applying the image segmentation theory in the image restoration field, we proposed a partition-based image restoration algorithm. According to image segmentation, edge detection, structural reconstruction and texture synthesis algorithm steps, we described algorithm's principles in detail, and gave the simulation experiment. Finally, according to the simulation experiment results, we summarized the talent and shortcoming of the algorithm.