Performance Analysis of Different Applications of Image Inpainting Based on Exemplar Technique

Abstract

In this age of rapidly developing image processing, inpainting has been a popular and practical art. Researchers have paid considerable attention to image inpainting throughout the years due to its enormous significance and effectiveness in a wide range of image processing applications, including the removal of scratches, the elimination of objects, and the modification of faces. It is one of the most challenging issues in image processing, demanding a comprehensive understanding of the image's texture and structure. The quality of inpainted image is a crucial factor which determines how close the inpainted image is to the original image. Many improvements have been implemented in the exemplar-based approach to increase the quality of inpainted regions containing structure and texture information. There are numerous ways to assess the quality of an inpainted image. In this study, the applications of exemplar based inpainting are evaluated using standard analytical measures including Sum of Absolute Difference (SAD), Peak Signal-to-Noise Ratio (PSNR), Correlation Coefficient, and Structural Similarity Index Measure (SSIM).