Effect of different traversal schemes in integral image coding

Abstract

Integral imaging (InIm) is a highly promising technique for the delivery of three-dimensional (3D) image content. During capturing, different views of an object are recorded as an array of elemental images (EIs), which form the integral image. High-resolution InIm requires sensors with increased resolution and produces huge amounts of highly correlated data. In an efficient encoding scheme for InIm compression both inter-EI and intra-EI correlations have to be properly exploited. We present an EI traversal scheme that maximizes the performance of InIm encoders by properly rearranging EIs to increase the intra-EI correlation of jointly coded EIs. This technique can be used to augment performance of both InIm specific and properly adapted general use encoder setups, used in InIm compression. An objective quality metric is also introduced for evaluating the effects of different traversal schemes on the encoder performance.

HEVC-based 3D holoscopic video coding using self-similarity compensated prediction

“…The RD performance is also compared to solutions where the 3D holoscopic image is decomposed into PVS (or multiview) content [19,23,24]. For a fair comparison, the following three coding solutions are tested:…”

“…x Pseudo video sequence coding: Alternatively, other compression methods propose to 'look' to the stack of micro-images, viewpoint images or ray-space images as a low resolution pseudo video sequence (PVS) and to encode it using a standard video encoder such as MPEG-2 [22], in [23,24], and H.264/AVC [25], in [19]. In these schemes, various scanning topologies to order the PVS are considered (raster, parallel, zigzag, spiral, and Hilbert).…”

“…In the sparse depth-dependent, depth information is estimated and an average PSNR is taken in a rendered view (with fixed angular position) over sets of pixels in different depth layers. In [24], a micro-image deviation-dependent metric is proposed. Based on the fact that the viewed 3D object is formed by the intersection of the light beams passing through neighboring microlenses, the average PSNR and standard deviation across all micro-images are calculated.…”

“…In this case the metric proposed in [24] is used, as given in (5) for all ‫ܯ‬ × ܰ micro-images of a 3D holoscopic image. As discussed in [24], the idea behind this metric is that the quality of the visualized 3D object by using a holoscopic display will be affected by drastic variations on the quality of neighboring micro-images.…”

“…As discussed in [24], the idea behind this metric is that the quality of the visualized 3D object by using a holoscopic display will be affected by drastic variations on the quality of neighboring micro-images. However, it should be also noted that the MIs contains only a small portion of the captured scene that will be never visualized directly, being difficult to correlate with a subjective quality assessment.…”