Research And Application Of Multi-core Architecture System Image Restoration Algorithm Based On Diffusion

The imgae inpainting is to restore the damaged image or remove the unwanted objects in the image. Currently, the technology has been widely applied to various fields, which mainly include the repair of medical images and repair scratches damaged photos, remove the objects in the image, restore the scene of the crime and soon.With the increasingly high resolution on the digital images, and the details that need to be addressed more and more complex, if we want to repair a high-resolution image, we could consume a very long time. We apply a stand-alone single-core-platform to deal with the entire repair process that is unacceptable. With the popularity and rapid development of multi-core platforms, based on multi-core architecture system parallelization in the digital image repair efficiency has been greatly improved.This paper introduces the research-situations at home and abroad about the technology of imgae inpainting firstly; then introduces the basics of the multi-core architectures and parallel computing; and then according to the existing algorithms, and we research the key topics of this paper:a new inpainting algorithm---diffusion inpainting algorithm.This algorithm is based on the reverse thinking of the seed filling algorithm. First, we select a point in the area that is a damaged area.Then the point as the center of the area launch many point-rays, and according to the information and color of existing point-rays, filled the basic information and color of the repairing point, and loops the program until the algorithm is completed. This algorithm can restore the color, and achieve a good effect.Finally, the algorithm is applied to the system of multi-core architecture and improved the running processing rate greatly.