Image Processing Reference
In-Depth Information
10.3 Prediction of Visual Discomfort Induced by Motion
It is well accepted that large disparity and large amount of motion are two main
causes of visual discomfort. To quantify this influence, it is necessary to design a
subjective experiment on visual discomfort induced by disparity and motion. In this
section, a well-controlled subjective experiment on 3D visual discomfort is intro-
duced as an example of how to design and conduct a reproducible and reliable
experiment in the area of 3D.
Three objectives are aimed at in the experiment. The first one is the comparative
analysis on the influence of different types of motion on visual discomfort, namely
static stereoscopic image, planar motion, and in-depth motion. The second one is
the investigation on the influence factors for each motion type, for example, the
disparity offset, the disparity amplitude, and velocity. The third one is to propose an
objective model for visual discomfort. In addition, the influence from viewers
3D
experience, i.e., the differences between experts and naive viewers is introduced.
'
10.3.1 State of the Art
As we already introduced in Sect.
10.2
, there are many possible factors that would
induce visual discomfort. Most of the causes have been well studied for the case of
still stereoscopic images. For stereoscopic 3D videos, as the only difference
between stereoscopic image and video is the motion, the influence of motion on
visual discomfort has been widely investigated recently.
In-depth motion is one of the significant factors that may cause visual discom-
fort. Studies already showed that visual discomfort increases with the in-depth
motion velocity [
9
,
11
,
12
,
31
,
33
]. However, the influence from disparity amplitude
(disparity range) and the disparity type (crossed or uncrossed) of in-depth motion
on visual discomfort are still under study. In [
11
], the results showed that disparity
amplitude of the moving object is not a main factor. However, in their recent study
[
12
] it is shown that visual discomfort increases with the disparity amplitude.
Furthermore, the results also showed that the in-depth motion with crossed disparity
would induce significantly more visual discomfort than the uncrossed and mixed
conditions. In [
33
], as they only analyzed the in-depth motion in the disparity range
of
1
with different velocities, there is no conclusion about the influence of
crossed or uncrossed disparity amplitude on visual discomfort.
The influence of the planar motion on visual discomfort was studied as well [
12
,
33
-
35
]. These studies showed high consistency on the conclusion that visual
discomfort increases with the motion velocity. However, the influence of
the disparity on visual discomfort led to different conclusions in these studies.
In [
33
], the results indicated that the disparity type, i.e., crossed and uncrossed
disparity, did not affect the visual discomfort thresholds. However, in [
1
], the
results showed that the crossed disparity will generate more visual discomfort
