Improved 3D integral imaging reconstruction with elemental image pixel rearrangement

Abstract

Computational reconstruction of integral imaging requires much more computational loads than optical reconstruction because of the adding and averaging of many elemental images digitally. Thus, to reduce the computational loads, the pixels of elemental images rearrangement technique (PERT) has been proposed. It can reconstruct a three-dimensional (3D) image very fast, but the size of the reconstructed 3D image is different from the conventional computational reconstruction due to no consideration of the empty space between pixels on the reconstruction plane. Therefore, in this paper, we propose PERT considering the projected empty space to correct the size of the reconstructed 3D images. To verify and support our proposed method, we carry out preliminary experiments and calculate the structural similarity index.