Evaluation of combined pre-processing and H.264-compression schemes for 3D integral images

Abstract Integral imaging is widely considered as a promising technique able to enhance video by adding scene depth, i.e. producing 3D video. However, an increase in spatial resolution of several orders of magnitude from todays 2D video is required to provide a sufficient depth fidelity. We have recently proposed a pre-processing and compression scheme for integral images showing good performance on a set of reference images. This scheme forms pseudo video sequences by transforming the static integral image into a set of orthoscopic projections, so called modified elementary images. The pseudo video sequence is then compressed using H.264 compression. In this paper we investigate how different orders of selecting the modified elementary images affect the schemes compression efficiency. Five different selection orders are evaluated on four reference images. The performance of each combination of selection order and reference image is presented. The obtained results indicate that for all but the most complex reference image, all selection order can provide a compression ratio of 320:1 or more, while keeping PSNR above 30 dB