Object removal and region filling in image

Sometimes people have been taken photographs, there has been the need to remove unwanted artefacts or damaged areas from these images. This is something previously conducted by experts, using computer tools and extensive user input to produce a visually realistic result. The difficulty of the problem lies in how to fill the missing regions, such that they are perceptually unnoticeable, whilst minimizing user input. Object removal and region filling algorithm proposed for removing unwanted objects from digital images. The challenge is to fill the hole that is left behind in visually plausible way. This problem has been addressed by two classes of algorithms: i) “Texture synthe-sis” for generating large image regions from sample textures, and ii) “In-painting” techniques for filling in small gaps. Each being used for images with different characteristics. This dissertation is concerned with investigating whether advances in texture synthesis techniques contain the necessary functionality to cross the divide between the two and successfully fill holes in a variety of images. The advantage of the proposed techniques is that it limits the amount of knowledge required by the user, and performs the region filling autonomously, once started