| Digital image inpainting is a technique which is to fill and inpaint the blank area based on the known information of image.Because image inpainting has been widely utilized in the field of archaeology?medical?industrial and film,the research of digital image inpainting is meaningful.The thesis researched the image inpainting algorithm for the damaged image,which both in subjective and objective evaluation can obtain the high quality image.The main contents of this paper are as follows:(1)This paper expounds the present situation of digital image inpainting technology at domestic and overseas and the background and meaning of digital image inpainting algorithms are studied.The partial differential equation technology,texture synthesis technique and partial differential equation and texture synthesis technique were analyzed.(2)The defect of traditional BSCB image inpainting algorithm in the inpianting speed,an improved BSCB image inpainting algorithm based on patch similarity was introduced;The TV image repair algorithm can easily create stair effect in smooth region of digital image,a self-adaptive image repair algorithm is proposed to improve the diffusion coefficient and adaptive positive parameters;Due to the defect of slow inpainting of CDD image inpainting algorithm,an adaptive CDD image inpainting algorithm is proposed.The experimental results show that the improved algorithm has been improved in performance and modification.(3)The inpainting algorithm of image based on partial differential and texture is used for the inpainting of geometry damage information and texture damage information;The main idea of this algorithm is to decompose the damage image into the geometric structure subgraph and the texture structure subgraph,the modified TV adaptive algorithm and the improved Criminisi algorithm are used to inpainting the damage image;The experimental results show that the modified TV adaptive algorithm and the improved Criminisi algorithm has high quality. |
PDF Full Text Request