Image Processing Reference
In-Depth Information
the total display resolution which is sampled for angular direction information.
Generally, the viewing resolution is simply defined by the number of micro-
lenses in the array. Hence, it is possible to enhance the viewing resolution by
decreasing the pitch of the micro-lens. However, a smaller micro-lens pitch leads
to a narrower viewing angle.
Therefore, an important requirement to improve one viewing parameter
without degrading the others (i.e., to provide 3D holoscopic content with good
visual quality) is to have acquisition and display devices with very high resolu-
tions. Consequently, efficient coding solutions become of cardinal significance
to deal with the large amount of information involved.
5.3
3D Holoscopic Video Representation Formats
A key requirement for successful 3D video applications is the choice of an efficient
representation in conjunction with an efficient coding scheme. In this sense, this
section describes three possible formats for representing the 3D holoscopic image
data:
micro-image based, viewpoint-image based,
and
ray-space-image based
.It
will be seen, in Sect.
5.5.1
, that a major factor for choosing a representation format
over another may be the compression efficiency it allows. Notice that, a depth-
enhanced representation of the 3D holoscopic content is also possible, but is not
addressed in this chapter. Moreover, to the best of authors
knowledge, depth-
enhanced 3D holoscopic representation and coding solutions have not been pro-
posed in the literature yet.
Furthermore, as discussed in Sect.
5.2.1
, one of the advantages of using a 3D
holoscopic imaging acquisition system is its high-level of display scalability, being
able, for instance, to generate 2D, stereoscopic 3D, and multiview content. Several
algorithms to generate 2D content from a 3D holoscopic image have been proposed
in the literature, mainly in the context of richer 2D image capturing systems [
6
].
In Sect.
5.3.4
, two of these algorithms, referred to as Basic Rendering and Weighted
Blending [
7
], are briefly reviewed, since they are used to generate the content for
each hierarchical layer of the scalable 3D holoscopic coding architecture described
later in Sect.
5.4
.
'
5.3.1 Micro-Image Based Representation
As a result of the used optical acquisition setup, where a micro-lens array is
overlapped with the image sensor, the 3D holoscopic image is inherently composed
of a 2D array of slightly different micro-images. Due to the different nature of the
3D holoscopic content, the first step to build an efficient coding scheme for this type
of content is to characterize their intrinsic correlations.
