Accelerated reconstruction of 3-D object images using estimated object area in backward computational integral imaging reconstruction

• DOI: 10.1007/3dres.04(2010)04
• Publication Date: 2011-01-25T07:39:33Z
• Authors: PeterByungHoon Han; Yongri Piao; Eun-Soo Kim
• Keywords: Integral Imaging

In this paper, we propose a novel approach for fast reconstruction of three-dimensional (3-D) object images by using the depth data of 3-D objects in the backward computational integral imaging reconstruction (B-CIIR) method. In the conventional B-CIIR method, once a reconstruction plane is selected, every pixel value composed of the reconstruction plane has to be calculated by averaging the pixel values all of the corresponding elemental images. In the proposed method, however, a potential object area to be displayed on the reconstruction plane is estimated in advance by using the depth data of the 3-D object. Here, the object's depth data can be easily obtained by applying a stereo matching algorithm to subimages transformed from elemental images. Then, contrary to the conventional B-CIIR algorithm, the number of image pixels to be calculated for reconstruction of 3-D image on the output plane can be reduced to that of the pixels within the potential object area. This might result in a dramatic reduction of calculation time and finally achieve to the much faster reconstruction of the 3-D object image than the conventional B-CIIR method. To show the feasibility of the proposed method, experiments are carried out and the results are presented.