FOR 3D DISPLAYS, LENS ACCOMMODATION IS VARIABLE

AND IT IS CONSISTENT WITH CONVERGENCE

Tomoki Shiomi

, Masaru Miyao

, Hiroki Hori

, Keita Uemoto

, Akira Hasegawa

Masako Omori

, Satoshi Hasegawa

, Hiromu Ishio

and Hiroki Takada

Nagoya University, Graduate School of Information Science, Chikusa-ku, Nagoya, Japan

Nagoya Bunri University Inazawa, Aichi, Japan

Kobe Women's University, Suma-ku, Kobe, Japan

Fukuyama City University, Fukuyama, Hiroshima, Japan

University of Fukui, Fukui, Japan

Keywords: Accommodation, Convergence, Simoultaneous Measurement, 3D Video Clips.

Abstract: Recently, 3D technology has been developing. It is generally explained to the public that, “During

stereoscopic vision, accommodation and convergence are mismatched and this is the main reason for the

visual fatigue caused by 3D”. The aim was to compare fixation distances between accommodation and

convergence in young subjects while they viewed 2D and 3D video clips. Measurements were made using

an original machine, and 2D and 3D video clips were presented using a liquid crystal shutter system. As

results, subjects’ accommodation and convergence were found to change the diopter value periodically

when viewing 3D images. These findings suggest that the ocular functions when viewing 3D images are

very similar to those during natural viewing. When subjects are young, accommodative power while

viewing 3D images is similar to the distance of convergence, and the two values of focusing distance are

synchronized with each other.

1 INTRODUCTION

Recently, 3D technology has been developing.

Today, 3D is not only used in movie theaters. Each

home appliance maker has started to sell 3D TVs

and 3D cameras.

It is generally explained to the public that,

“During stereoscopic vision, accommodation and

convergence are mismatched and this is the main

reason for the visual fatigue caused by 3D. During

stereoscopic vision, while accommodation is fixed

on the display that shows the 3D image,

convergence of left and right eyes crosses at the

location of the stereoimage”. Studies by Wann et al.

and Yano et al. (Wann, 1995); (Yano, 2004) found

that in natural vision lens accommodation is

consistent with convergence; that is, accommodation

and convergence matched. However, they noted that

fatigue might occur after extensive viewing of 3D

images because accommodation and convergence

are not matched when viewing these images.

According to the findings presented in our previous

report (Miyao, 1996), however, such explanations

are mistaken. We found that lens accommodation for

3D images is in fact consistent with convergence

among young subjects. However, our research has

not been recognized in the world. This may be

because the experimental evidence obtained in our

previous studies, where we did not measure

accommodation and convergence simultaneously,

was not strong enough to convince people. We

therefore developed a new device that can

simultaneously measure accommodation and

convergence.

The aim was to compare fixation distances

between accommodation and convergence in young

subjects while they viewed 2D and 3D video clips.

2 METHOD

In this experiment, the subjects were six healthy,

young men and women in their twenties (two had

uncorrected vision and four used soft contact

780

Shiomi T., Miyao M., Hori H., Uemoto K., Hasegawa A., Omori M., Hasegawa S., Ishio H. and Takada H..

FOR 3D DISPLAYS, LENS ACCOMMODATION IS VARIABLE AND IT IS CONSISTENT WITH CONVERGENCE.

DOI: 10.5220/0003908407800783

In Proceedings of the International Conference on Computer Graphics Theory and Applications (IVAPP-2012), pages 780-783

ISBN: 978-989-8565-02-0

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

lenses). We obtained informed consent from all the

subjects and the approval from the Ethical Review

Board of the Graduate School of Information

Science at Nagoya University.

We placed an LCD monitor facing the subjects at

a distance of 1m from them. We presented either a

2D or a 3D video clip on the monitor; in both

images, a spherical object moved forward and

backward with a cycle of 10 seconds (Figure 1). The

spherical object appeared as a 3D video clip located

at a virtual distance of 1m (i.e., the location of the

LCD monitor) and moved toward the subjects to a

virtual distance of 0.35m in front of them. We asked

the subjects to gaze at the center of the spherical

object for 40 s and measured their lens

accommodation and convergence distance during

that time. The 3D video clip was presented using a

liquid crystal shutter system and a circular polarizing

filter system. The 2D video clip was presented using

only a liquid crystal shutter system.

Figure 1: Spherical object video clips.

We developed an original machine by combining

WAM-5500

and EMR-9

to perform these

simultaneous measurements

Figure 2: WAM-5500.

WAM-5500 (Figure 2) is an auto refractometer

(Grand Seiko Co., Ltd.) that can measure

accommodative power under natural conditions for

the case in which both eyes are open. It can

continuously record accommodative focus distance

at a rate of 5 Hz, thus achieving reliable and accurate

measurements of accommodation.

EMR-9 (Figure 3) is an eye mark recorder (NAC

Image Technology Inc.) that can measure the

convergence distance using the pupillary/corneal

reflex method. Its resolution for eye movement is

0.1 degree, with a measurement range of 40 degrees

and sampling rate of 60 Hz. The convergent focus

distance can be easily calculated from the obtained

binocular eye movement data.

Figure 3: Eye mark recorder EMR-9.

We used a liquid crystal shutter system or a

circular polarizing filter system combined with the

respective binocular vision systems to present 2D

and 3D video clips. The experimental environment

is shown in Figure 4.

Figure 4: Overview of the experiment.

The video clips we used in the experiment are

trademarked as Power 3D

video clip (Olympus

Visual Communications, Corp.). Power 3D is an

image creation technique that combines near and far

views in a virtual space and has multiple sets of

virtual displays whose positions can be adjusted.

Power 3D presents a video clip that is similar to a

natural image.

3 RESULT

The measurements for the six subjects showed

FOR 3D DISPLAYS, LENS ACCOMMODATION IS VARIABLE AND IT IS CONSISTENT WITH CONVERGENCE

781

roughly similar results. For 3D vision, the results of

2D and 3D for one subject are shown respectively in

Figs. 5 and 6, as typical examples.

Figure 5: 3D liquid crystal shutter system (age:23, male).

Figure 6: 2D liquid crystal shutter system (age:23, male).

When Subject (23-year-old male wearing soft

contact lenses) viewed the 3D video clip presented

using the liquid crystal shutter system,

accommodation varied between approximately 1.0

diopter (100cm) and 2.5 diopters (40cm), whereas

convergence varied between approximately 1.0

diopter (100cm) and 2.7 diopters (37cm). The

changes in the respective diopter values had almost

the same amplitude and were in phase, fluctuating

synchronously with a cycle of 10 s, which

corresponded with the cycle of the 3D video clip

movement.

When the subject was viewing the 2D video clip,

the diopter values for both accommodation and

convergence remained almost constant at around 1

diopter (1m).

The mean values of accommodation and

convergence for the six subjects viewing the 2D

video clip were 0.96 ± 0.12 and 0.96 ± 0.07,

respectively. The difference between

accommodation and convergence was negligible.

When the subjects were viewing the 3D video clip,

the values of accommodation and convergence were

1.29 ± 0.11 and 1.32 ± 0.08, respectively. The

difference between accommodation and

convergence in this case was approximately 0.03

diopters, which is also negligible. Therefore, we can

say that there is not much quantitative difference in

the fixation distances between accommodation and

convergence when the subject views either the 2D or

3D video clip.

4 DISCUSSION

Wann et al. stated that within a virtual reality

system, the eyes of a subject must maintain

accommodation at the fixed LCD screen, despite the

presence of disparity cues that necessitate

convergence eye movements to capture the virtual

scene. Moreover, Hong et al. stated that the natural

coupling of eye accommodation and convergence

while viewing a real-world scene is broken when

viewing stereoscopic displays (Hong, 2010).

In addition to the above two, Hoffman et al. and

Ukai et al. (Hoffman, 2008); (Ukai, 2008) stated that

if there is inconsistency between accommodation

and convergence, then accommodation to a 3D

object is fixed at the position of the display. In this

study, however, the result showed good

synchronization between accommodation and

convergence during stereoscopic vision and this did

not occur. This suggests that the difference between

accommodation and convergence is probably not the

main reason for visual fatigue, motion sickness, and

other problems.

We can also say that the kind of results presented

herein could be obtained because the 3D images

used in the experiments were produced not by

conventional means but with Power 3D, whose

images are extremely close to natural viewing.

In fact, conventional 3D and the Power 3D on

HMD have been compared experimentally in our

previous study (Hasegawa, 2009). This study found

that the result of Power 3D is closer to natural vision

than that of conventional 3D.

Therefore, we consider that as long as 3D images

are made using a proper method, accommodation

and convergence should almost always coincide, and

that we can view such images more easily and

naturally.

In conventional “accommodation-convergence

discrepancy theory,” accommodation is fixed on the

display during virtual 3D vision, although focus of

convergence is consistent with the location of the

virtual image.

IVAPP 2012 - International Conference on Information Visualization Theory and Applications

782

However, our present experiment findings

suggest that accommodative focus is nearly

consistent with the location of the stereoimage.

There is also an opinion that an image is seen as

blurred if the accommodative focus is not on the

display but on the virtual position. In this experiment,

however, all subjects said that the image was clear.

Figure 7:

Naked vision for primary school child in

1998-2009.

The blue solid line indicates naked vision value

corresponding to refractive value. The X axis

indicates the refractive value, the left Y axis

indicates naked vision value measured in geometric

mean, and the right Y axis indicates LogMAR

values. LogMAR is the visual acuity log

transformed.

Let us consider at the case of virtual 3D images

popping forward, for example the case in Fig. 5 in

which an LCD monitor was placed 100 cm in front

of the subjects and then a virtual spherical object

moved to 40 cm in front of subjects. The theoretical

blurring that occurs with virtual 3D images is

approximately equal to that of a subject who has a

myopic view of infinity of -1.5 diopters (nearly

equal to >5.0 m). That is blurring of far visual acuity

such as in a subject with myopia of -1.5 diopters.

The Nagoya City Education Committee has a

statistics on myopic children and visual acuity with

no astigmatism (Fig. 7). This data is the result of

detailed examination of thousands children (11 years

old) in Nagoya city (Japan) by more than 100

ophthalmologists. According to Fig. 7, among these

children visual acuity of -1.5 diopters is about the

geometric mean value of 0.35 (LogMAR value is

about -0.46). In addition, blurring is also greatly

affected by pupil diameter (Smith, 1991). For bright

virtual 3D images, pupil diameter contraction and

focal depth becomes deeper. Therefore, blurring is

reduced. In fact, the subjects in present experiments

did not recognize blurring in the 3D video clips.

5 CONCLUSIONS

In this experiment, we simultaneously measured

accommodation and convergence for subjects

viewing 2D and 3D video clips. The difference in

the eye movements for accommodation and

convergence is equally small in the cases of the

observation of both 2D and 3D video clips. This

suggests that the difference between accommodation

and convergence is probably not the main reason for

visual fatigue, motion sickness, and other problems.

In the future works, we are going to investigate

subjects’ visual information, for example, whether

subjects see a blurred image, and so on. In this

experiment, subjects didn’t accommodate on the

display in gazing 3D video clip. However there is no

report that subjects could not see a video clip.

Therefore, it is very important to investigate more

accurately the information which subjects get when

viewing 3D vision.

REFERENCES

J. P. Wann, S. Rushton, M. Mon-Williams, 1995. Natural

problems for stereoscopic depth perception in virtual

environments, Vision Res, 35, 19, 2731-2736.

S. Yano, M. Emoto, T. Mitsuhashi, 2004. Two factors in

visual fatigue caused by stereoscopic HDTV images,

Displays, 25, 4, 141-150.

M. Miyao, S. Ishihara, S. Saito, et al., 1996. Visual

Accommodation and Subject Performance during a

Stereographic Object Task Using Liquid Crystal

Shutters, Ergonomics, 39, 11, 1294-1309.

Hong H and Sheng L, 2010. Correct Focus Cues in

Stereoscopic Displays Improve 3D Depth Perception,

SPIE, Newsroom.

D. M. Hoffman, A. R. Girshick, K. Akeley, and M. S.

Banks, 2008. Vergence-accommodation conflicts

hinder visual performance and cause visual fatigue,

Journal of Vision 8, 33, 1-30.

K. Ukai, P. A. Howarth, 2008. Visual fatigue caused by

viewing stereoscopic motion images: background,

theories, and observation, Displays 29, 2, 106-116.

S. Hasegawa, M. Omori, T. Watanabe, K. Fujikake, and M

Miyao, 2009, Len Accommodation to the Stereoscopic

Vision on HMD, Lecture Note in Computer Science,

5622, 439-444

G. Smith, 1991. Relation between spherical refractive

error and visual acuity, Optometry and Vision Science,

68, 8, 591-598.

FOR 3D DISPLAYS, LENS ACCOMMODATION IS VARIABLE AND IT IS CONSISTENT WITH CONVERGENCE

783