查看文章

Light fields capture a large number of samples of light rays in both intensity and direction terms, which allow post-processing applications such as refocusing, shifting view-point and depth estimation. However, they are represented by huge amount of data and require a high-efficient coding scheme for its compression. In this paper, light field raw image data is decomposed into multi-views and used as a pseudo-sequence input for state-of-the-art codecs such as High Efficiency Video Coding (HEVC). In order to better exploit redundancy between neighboring views and decrease distances between current view and its references instead of using conventional orders, views are divided into four smaller regions and each region is scanned by a snake order. Furthermore, according to this ordering, an appropriate referencing structure is defined that only selects adjacent views as references. Simulation results show that …