Introducing a new fast exemplar-based inpainting algorithm

Inpainting is the process of reconstructing damaged parts of an image in a visually plausible way. Recently, digital image inpainting has been an active field of research in the area of image processing. Meanwhile, exemplar-based methods have had a major contribution in developing this field of research. The main idea in these methods is based on copy-and-paste texture synthesis. The time-complexity of these methods are rather high. So, many different approaches were introduced to tackle this problem. This paper presents a new algorithm for reconstructing missing parts of an image based on exemplar-matching techniques in which both performance and speed of the algorithm increases. Experimental results confirm the effectiveness of the proposed method.