Page 118 - ITU Journal, ICT Discoveries, Volume 3, No. 1, June 2020 Special issue: The future of video and immersive media

ITU Journal: ICT Discoveries, Vol. 3(1), June 2020

4.5.3 Subjective testing protocols is the first international standardization activity targeting
digital hologram compression.
Despite many years of effort, current point cloud objective
quality metrics do not correlate well with subjective judg- The JPEG Pleno Holography initiative has to solve a num-
ments [24]. The best judge of point cloud visual quality ber of problems that need to be addressed before stan-
continues to be human visual appraisal. The JPEG Com- dardization can effectively be launched.
mittee has adapted an existing subjective testing proto-
col [24] for the evaluation of point cloud quality. This pro- First, several test data have been collected for exper-
tocol involves the following steps: iments. These holograms were optically captured via
interferometry or synthesized via Computer-Generated
Holography (CGH) techniques to cover microscopic (holo-
• A view path is determined for each reference point graphic microscopy and tomography) and macroscopic
cloud in the form of a smooth path of movement applications (display and printing). Several hologram
around the point cloud that allows the observer to databases [27] are currently available via the JPEG web-
view the point cloud from a variety of viewpoints ap- site.
propriate for the content.
Second, holograms are typically represented as complex-
• For each point cloud, a 2D video is created where the valued data. Hence, codecs have to account for this char-
frames are the projections of the point cloud onto acteristic and real & imaginary or amplitude & phase rep-
a virtual camera image plane as it moves along the resentations need to be supported.
complete view path. Each video lasts for a duration
of 24 seconds at a rate of at least 30 frames per sec- Third, since no advanced holographic display exists cur-
ond. rently, alternative quality assessment procedures need to
be designed to evaluate the decoded holograms [28].
• The videos of the reference and the processed point
To address these issues, the JPEG Committee has been
cloud are presented side-by-side, synchronized, on a
2D monitor. Subjects are asked to judge the degree conducting exploration studies aiming at (1) defining ob-
of impairment of the processed point cloud relative jective and subjective quality evaluation procedures for
to the reference point cloud on a 5 point degrada- holography and (2) designing a numerical reconstruction
tion/impairment scale according to the established software to be used as a tool for hologram reconstruction.
Double Stimulus Impairment Scale (DSIS) system Currently, the common test conditions (CTC) are being de-
outlined in Recommendation ITU-R BT.500.13 [25]. fined and a call for proposals (CfP) is scheduled for 2020.

• Collected data has outliers removed and is averaged 4.6.2 Numerical reconstruction software
into mean ppinion scores according to Recommen-
dation ITU-R BT.500.13 [25]. Before discussing the test procedures, this section ad-
dresses the numerical reconstruction software that was
designed to facilitate testing. Indeed, to assess the per-
The point cloud activity issued a Call for Evidence on Point formance of holographic data coding solutions, it is neces-
Cloud Coding in early 2020 and next steps for this activity sary to compare the visual quality and/or signal-to-noise
are to evaluate submissions to the call and progress to a ratios of scene images numerically reconstructed from
Call for Proposals for static point cloud coding solutions the decoded and original reference holograms.
that may form a new part in the JPEG Pleno standard.
To this end, a Numerical Reconstruction Software for
Holography (NRSH) has been developed in Matlab, en-
4.6 JPEG Pleno Holography
abling the reconstruction of different types of holograms,
4.6.1 Current status including complex and real-valued, monochromatic and
colourful holograms. Utilizing one unique well-validated
Holograms deliver a realistic three-dimensional view- software environment for this process guarantees that no
ing perception with no vergence-accommodation conflict unexpected distortions are introduced, which might be
(VAC), since they record and reproduce 3D scenes by rep- the case when using proprietary software.
resenting both the amplitude and phase of light. Recently,
digital holography has received considerable attention Fig. 11 shows the reconstruction of a complex-valued
and has become popular in digital microscopy and inter- hologram from two different viewing angles using NRSH
ferometry. Moreover, holographic displays and printing software. Depending on the type of input hologram,
are attracting increasingly more interest. However, since three methods can be used to simulate the propagation
the quality of digital holograms depends on their pixel of light waves: the Angular Spectrum, Fresnel and Fraun-
pitch and resolution [26], they contain a massive amount hofer diffraction formulas. Furthermore, to reconstruct
of data. Hence, efficient compression is necessary to real- the hologram from different viewing angles, it is possible
ize holographic imaging services. JPEG Pleno Holography to simulate a synthetic pupil aperture in the observer’s

96 © International Telecommunication Union, 2020