Quality Assessment - 3DTV: Processing and Transmission of 3D Video Signals - page 162

6.3 3D Video Quality Evaluation Methods

Due to the multi-dimensional attributes of 3D video such as image quality,
depth quality and visual comfort, it is currently difficult to evaluate the
perceptual quality of 3D applications, services and the suitability of 3D video
coding algorithms for compression and transmission, without resorting to
full subjective tests. Therefore subjective test campaigns are widely carried
out to understand the true quality of 3D representation [3]. Subjective quality
evaluation methodologies for stereoscopic television pictures are described in
ITU-R BT.1438 [10]. Subjective evaluation procedures in this recommendation
are based on the ITU-R BT.500.11 quality assessment recommendation for
television pictures [11]. The main assessment methodologies considered in
[10] and [11] are described in Table 6.1.
DSCQS and SSCQS methods are used for subjective assessment of 3D
video, as these methods are recommended by standardization bodies for
3D video quality measurements and are widely used in 3D video research
[12, 13]. The presentation structure of test materials and the quality-rating
scales for both DSCQS and SSCQS methods are illustrated in Figure 6.8
and Figure 6.9 respectively. In the DSCQS method, reference and processed

Table 6.1 Assessment methodologies for subjective testing

Assessment methodology Description

Double-Stimulus-Continuous-
Quality-Scale (DSCQS)
method

Measures the quality of the processed (or
distorted) video sequence compared to its
undistorted reference (or original video
sequence). The assessor provides a quality
index for both processed and undistorted
video sequences.

Single-Stimulus-Continuous-
Quality-Scale (SSCQS)
method

A single video is presented without its
reference and the assessor provides a quality
index individually for each video sequence
in the stimulus set.

Threshold estimation method Series of video sequences are presented

sequentially in time. The assessor is asked to
assess the point at which impairment
becomes visible.

Stimulus comparison method Series of video sequences are presented

sequentially in time. The assessor is asked to
assign a relation between two consecutive
video sequences.