Object removal and inpainting technique by Lattice Boltzmann method and exemplar method in color image

Abstract

In this paper we use image inpainting technique in a color image. Inpainting is a technique in which, a missing area or pixels are replaced by adequate neighbouring pixels so that the resultant image obtained will look as no defect is there. In this paper we have removed an object from the color image and the vacant pixels are filled with the help of the surrounding neighbouring pixels. Two techniques are used here for removing and inpainting an object from the still color image, lattice boltzmann and exemplar method. These techniques were selected as Exemplar method of inpainting was used in many algorithms and have given a good result and lattice Boltzmann was usually used in experiments conducted in fluid dynamics. As motion of pixels can be compared to motion of fluids, lattice boltzmann method gives a better match for filling in the regions. Different variables for the inpainted image was calculated for both the methods. The signal to noise ratio and self-similarity index of the image is calculated for both inpainting techniques. According to the signal to noise ratio and the time taken for inpainting, it is found that lattice Boltzmann method of inpainting gives a better inpainted result..