Improved Image Inpainting Using Hybrid Method Consisting Of Convolution and Exemplar Based Algorithm

Image Inpainting is the art of filling in missing data in an image. The purpose of inpainting is to reconstruct missing regions in a visually plausible manner so that it seems reasonable to the human eye. There have been several approaches proposed for the same. The objective of this dissertation work is to review two such methods – Convolution and Exemplar based Algorithm and combines these into a Hybrid technique. Convolution based image inpainting algorithms are very rapid, however in many cases; they don't provide adequate results in sharp details such as edges. Exemplar-based methods provide impressive results in recovering textures and repetitive structures. However, their ability to recreate the geometry without any example is limited and not well understood. The aim is to incorporate such enhanced characteristics in the Hybrid technique which will be better than the individual techniques. This paper summarizes the combination of these algorithms and proposes a new Hybrid Inpainting Algorithm for inpainting large as well as small regions in less time. Future aspects of the study are also discussed.