Light Field Image Coding with Disparity Correlation Based Prediction

Plenoptic camera has emerged as an advanced optical device to record both direction and intensity of rays passing through each point in space. Its productions called light field (LF) images or lenslet images contain more abundant visual information, making it necessary to reduce the considerable redundancies to facilitate storage and transmission. A novel light field image coding scheme is proposed in this work which employs a disparity correlation based prediction structure. With the proposed prediction scheme, LF images are partitioned into sub-aperture images to form a synthesized sequence, and then coded by the high efficiency video coding (HEVC) standard and its multi-view extension (MV-HEVC). The experimental results demonstrate that the proposed disparity correlation based prediction scheme is effective and robust for various scenes.