IJIRST –International Journal for Innovative Research in Science & Technology| Volume 2 | Issue 11 | April 2016 ISSN (online): 2349-6010 All rights reserved by www.ijirst.org 336 Object Removal and Region Filling based on Exemplar-based Method Pritam N. Shet Prof. Kimberly Morais UG Student Professor Department of Electronics and Telecommunication Engineering Department of Electronics and Telecommunication Engineering Don Bosco College of Engineering, Fatorda-Margao, Goa Don Bosco College of Engineering, Fatorda-Margao, Goa Vivek Tendulkar Vivek Singh UG Student UG Student Department of Electronics and Telecommunication Engineering Department of Electronics and Telecommunication Engineering Don Bosco College of Engineering, Fatorda-Margao, Goa Don Bosco College of Engineering, Fatorda-Margao, Goa Vyshakh. V UG Student Department of Electronics and Telecommunication Engineering Don Bosco College of Engineering, Fatorda-Margao, Goa Abstract Image Inpainting is a technique used to restore the lost parts of an image and rectify any alterations in the original image in visually appealing manner. Inpainting technique focusses upon getting any modified or damaged image back to its original form. This technique has got a list of applications including removal of occlusions, such as stamps, subtitles, text on photographs, rebuilding of damaged photographs and films, removal of unwanted objects or blur and red eye correction. In this paper we provide a brief review of Exemplar based image Inpainting technique. Keywords: Inpainting, Exemplar based method, Texture synthesis, Distance matrix, patch priorities _______________________________________________________________________________________________________ I. INTRODUCTION Restoration of missing or damaged portions of an image is very important practice used extensively in artwork restoration. This very technique is known as inpainting or retouching. This practice consists of filling in the missing areas of any old picture degraded over ages or modifying the damaged pictures in non-detectable manner [1]. In addition to common inpainting applications like restoration of photographs films and paintings, removal of superimposed objects or characters such as text, subtitles, stamps and publicity from images, inpainting can also be used to produce special effects. Traditionally, skilled artists and studio professionals have to perform image inpainting manually using popular licensed tools such as Photoshop. Firstly, Bertalmio et al [1,2] had introduced algorithm for digital inpainting of still images that produced very impressive results exactly fitting into the definition of inpainting. This algorithm, however, required several minutes on windows personal computers for the inpainting of relatively small areas. From this point, extensive work on inpainting has been carried out across the globe and still counting on developing algorithms capable of producing similar results in just a few seconds. A very effective algorithm that gets rid of this unacceptable time consumption for inpainting of small areas is Exemplar based inpainting method. This method relies upon a simpler and faster algorithm and the results produced by this algorithm are comparable to those found in the literature, but two to three times faster in speed. We illustrate the effectiveness of our approach with examples of reconstructed photographs and vandalized images. II. EXEMPLAR-BASED METHOD Based on seminal work on texture synthesis [2,3], a family of inpainting methods has appeared in the last decade with the aim of better recovering the texture of the missing area rather than simply replacing a patch with any matching exemplar. The goal of texture synthesis also referred to as sample-based texture synthesis is to create a texture from a specified background texture sample in such a way that, the produced texture is with a similar visual appearance as that of the background. Exemplar-based inpainting employed for retouching a large part, is inspired by local region-growing methods that grow the texture pixel-by-pixel or patch- by-patch at a rate of one pixel or one patch at a time, while maintaining coherence with nearby pixels or patches. Local pixel-based texture synthesis techniques widely rely on Markov random fields (MRFs) modelling of textures [3,4]. But in our algorithm we propose to tackle this modelling technique which runs very complex graphical approaches to find information of missing pixels, by running algorithms that exploit both locality (the colour and data of a pixel being assumed to depend on its neighboring pixels) and stationarity (the best replaceable local pixel is independent of the pixel location), the missing pixels are learned by sampling