DIGITAL IMAGE AND VIDEO INPAINTING

Abstract

Image Inpainting is a synonym for Image Interpolation. It is a technique of modifying an image in an undetectable form. Inpainting involves filling in part of an image or video using information from the surrounding area. It has many profound applications in image processing, from the. restoration of damaged photographs and movies to the removal of selected objects from images and videos. In this dissertation, we introduce a novel method for filling-in the missing information in an image. The technique presented here is able to reconstruct the structure as well as texture in images. - The inpainting is done in such a manner, that it fills the damaged region or "holes" in an image, with surrounding color and texture information. The algorithm is based on solving Laplace equation with Dirichlet boundary conditions, incorporated with texture filling, and its main asset is the bounded search window. The filling in the missing area is done in the order from outside to inside depending on the priority, which is calculated from the confidence value of the pixels. Our technique produces very good results in very less time. In this dissertation work, we also propose a technique for Video Inpainting. The approach involves the automated tracking of the object selected for removal, followed by filling-in the holes while enforcing the global spatio-temporal consistency. The tracking is based on color-based and motion-based segmentation of -objects. For reconstructing the videos, we have extended our image inpainting algorithm that enforces spatio-temporal consistency. A number of examples on videos sliow the effectiveness of our algorithm in tracking and removing large occluding objects as well as thin scratches. Video Inpainting based restoration is useful for a variety of tasks, including, but not limited to sophisticated video removal of undesired static or dynamic objects, correction of missing/corrupted video frames in movies and restoration of old scratched films. The implementation of programs is done using MATLAB version 6.5 and Windows XP operating system.