Epipolar Plane Image-Based Lossless and Near-Lossless Light Field Compression

Abstract

Recent advances in the computer technology has enabled many theoretical ideas in computer vision to become practical also within camera technology. Light field technology aims at replacing the traditional camera design by acquiring the intensity of light rays to realize advanced applications in media and medical industries. The abundant amount of information poses high processing, transmission and storage requirements. Therefore, a need for efficient compression has gained researchers attention in this area. We present EPIC, an efficient lossless compression scheme for light field images which extends CALIC by exploiting light field structure. We have shown that the proposed scheme achieves around 6% improvement in compression performance relative to HEVC-lossless anchor with a significant reduction in the encoding time. We have also extended the proposed method for the near-lossless case which achieves similar rate-distortion performance as HEVC-lossy anchor in addition to a high encoding efficiency and a guaranteed-error bound.