Automatic 2D-to-3D image conversion using KNN search

Last few years there are many techniques available that convert 2D image to 3D ones. Methods which make use of human operator have been very useful but they required more time and costly too, known as semiautomatic method. So to remove this drawback developed automatic 2D-to-3D image conversion method which does not required human operator's help and it also reduces complexity. But these methods still have not able to produce good quality image due to poor quality of depth image used for image conversion. The main step of 2D-to-3D image conversion is depth estimation of given 2D image. So here we introduced a novel technique to convert 2D image to 3D stereoscopic image with better depth image. The proposed method is based on estimating depth image of our query 2D image directly from database which contains many 2D images with their depth image by applying nearest neighbor search.