The improvement of the robust exemplar-based inpainting algorithm for SONAR images

The goal of digital inpainting algorithms is to automatically fill-in the missing areas or to modify the damaged ones, making them as close as possible to the original image. This paper deals with the improvement of the robust exemplar-based inpainting algorithm, in case of real SONAR images. This task is accomplished by selecting the weight α of the robust priority function. The performance of the improvement procedure is evaluated in terms of the visual quality and the Peak Signal to Noise Ratio (PSNR) of the inpainted image.