Research On The Image Inpainting Methods Based On Sample Block

Image inpainting is a popular research direction in the field of image processing and has a wide range of practical value.By establishing an image inpainting model,the damaged area is simulated and created to ensure the integrity of the image information in the process of the image inpainting.The technology can be used in the recovery of art works,the simulation of criminal investigation evidence,the reconstruction of medical images,daily matting and the repair of old videos.In this thesis,the method of image inpainting based on sample block is studied.This thesis introduces the background and research status of image restoration,and analyzes the shortcomings and limitations of traditional image inpainting methods.On this basis,this thesis mainly completes the following two aspects of work:1.Improved image inpainting method is improved based on the priority function.In view of the situation that the confidence term may plunge to zero in the later stage of image inpainting,the expression of the priority function is improved,the gradient factor and the equilibrium coefficient are introduced to effectively restrain the reduction range of the priority function.In addition,combining with the L1 distance matching criterion,the matching accuracy of the sample block is increased.2.Image inpainting method based on adaptive adjustment of the sample block size.In the traditional image inpainting methods based on the sample block,the template size of sample block is fixed,which cannot flexibly adapt to the image with complex structure information.Combining with the improved principle of structural similarity,the size of the sample block can be adjusted adaptively according to the complexity of the structure of the edge of the image to be repaired,thus the accuracy and efficiency of the image inpainting is ensured.The LoG edge detection method is introduced to enhance the recognition ability of the model for the damaged image edge,which effectively avoids the incomplete repairing of the damaged area.In this thesis,the experimental simulation is carried out,which shows the effectiveness and feasibility of the proposed inpainting methods.