Image Inpainting Algorithm Based On Patch Structure Sparsity

Image inpainting refers to the process of restoring missing or damaged areas in an image. The purpose of image inpainting might be to restore damaged portions of an image which is visible to human eyes. Over recent years, this technique has been used in many areas such as restoring image from scratches or text overlays, loss concealment in a context of impaired image transmission, object removal in a context of editing and so on. Due to the wide spread application of this technique, so the research of image inpainting has great significance on theory and reality.This paper focused on Criminisi algorithm, the research is mainly focused on it Advantages and disadvantages and present an improved algorithm based on patch structure sparsity. This algorithm calculates the priority by using patch structure sparsity. It makes the inpainting order more reasonable. Calculating the patch window size base on patch structure sparsity, improved the flexibility of inpainting. We introduce color histogram when matching the sample block and make the match more accurate. Finally, we replace global search to local search to reduce time cost.the imperfection of Criminisi algorithm are revised by the test. It can restore different damaged image. Moreover, with compared to Criminisi algorithm, the inpainted time has reduced 49.94 seconds on average, PSNR has improved 3~4dB, SSIM has increased 0. 0030 on average.