Survey on measures of image similarity applied in image inpainting base on deep learning

Abstract

The research of image similarity measures is an important content of the computer image processing. It have been widely studied and applied in image inpainting, image super resolution, image matching and other applications. In these application, the choice of image similarity measure has a direct impact on the final results. In order to enable more researchers to explore the image similarity theory and its specific role in different applications, this paper summarizes the research status in this field. This paper first classifies the existing image similarity evaluation methods from two different perspectives. Next, since the performance of image similarity measure is closely related to its application field, so we select the image inpainting field based on depth learning, and illustrate the application of image similarity measurement methods in this field. Then we roughly compare the advantage and disadvantage of these methods. Finally, we gave our conclusions and prospects.