AUTOMATIC SHAKE TO ENHANCE

FRASER-WILCOX ILLUSIONS

Kazuhisa Yanaka, Ryuto Mitsuhashi and Teluhiko Hilano

Kanagawa Institute of Technology, 1030 Shimo-ogino, Atsugi-shi, Kanagawa-ken, 243–0292, Japan

Keywords: Optical illusion, Fraser-wilcox illusion, Illusory motion, Peripheral drift illusion.

Abstract: The Fraser-Wilcox illusion, which is an optical illusion first found by Fraser and Wilcox in 1979 and later

classified into the peripheral drift illusion that was presented in 1997, is the illusion that a disk drawn on a

still image looks as if it is rotating. Recently, Kitaoka proposed an optimized Fraser-Wilcox illusion type V

in which a stronger illusion can be perceived. This proposal has attracted a great deal of attention, but not

everyone can see the illusion. It is well known that the effect of some existing illusions is reinforced by

shaking the image by hand, and we therefore developed a system in which a still image displayed on the

screen of an ordinary PC can be shaken automatically by using our software. Experimental results

demonstrated that the strength of some types of Fraser-Wilcox illusions can be enhanced considerably by

using the proposed system.

1 INTRODUCTION

Optical illusions are thought to be a kind of

malfunction that occurs when the highly

sophisticated human visual system processes retinal

images. The study of optical illusions is therefore

useful in terms of understanding human cognitive

mechanisms.

Among the various illusions that have been

reported up to now, there is a very interesting one,

usually called the illusory motion or the self-

animated image, in which parts of a still picture look

as if they are moving just by glimpsing it. A typical

example is the Fraser-Wilcox illusion (Fraser and

Wilcox, 1979). In this illusion, disks that consist of

repeated asymmetric patterns are drawn on paper or

displayed on a screen and appear to rotate

spontaneously, even though in actuality the image is

perfectly stationary. The illusion is strong enough

for those who are accustomed to it, but it is so faint

that beginners of illusion can hardly see it.

It is empirically known that the strength of an

optical illusion, not restricted to the Fraser-Wilcox

one, can sometimes be reinforced when the figure is

shaken: for example, by observers repeatedly

moving their glasses up and down. However,

reproducibility cannot be expected with such a

method. We could find no reports on how to shake

or with what kind of illusions this method is

effective. Therefore, we developed a system that

automatically shakes an illusion picture displayed on

a PC screen and then studied this phenomenon

experimentally.

2 FRASER-WILCOX ILLUSION

After the discovery of the Fraser-Wilcox illusion, a

similar illusion was reported by Faubert and Herbert

(1997). They called the illusion the peripheral drift

illusion (PDI) because it is caused by peripheral

vision and the illusion disappears when the pattern is

gazed at directly. Since then, the Fraser-Wilcox

illusion has been thought to be a typical example of

PDI. Many researchers have studied it intensively in

an attempt to reveal the mechanism behind such

illusions. One possibility is that the gradient of the

pattern that is scanned by the fixation movement of

the eye cheats the neurons that detect movement. An

interesting theory has recently been presented that

explains illusionary motion by combining a time

filter and a spatial filter (Fermüller, Ji, Kitaoka,

2010). Although this theory is feasible, further

research would be necessary to develop any final

conclusion.

In contrast, intensive experimental studies have

been done to clarify the conditions in which a

stronger illusion is perceived. A representative

405

Yanaka K., Mitsuhashi R. and Hilano T..

AUTOMATIC SHAKE TO ENHANCE FRASER-WILCOX ILLUSIONS .

DOI: 10.5220/0003321904050408

In Proceedings of the International Conference on Computer Vision Theory and Applications (VISAPP-2011), pages 405-408

ISBN: 978-989-8425-47-8

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

example of such studies is Kitaoka’s work. He not

only created a number of novel, beautiful, and strong

“optimized” Fraser-Wilcox illusions, he also

classified the illusions into six types based on the

underlying rules. Among them, Type V (Kitaoka, A.,

2009) is unique because the effect is very strong

compared to other types. In addition, in Type V

color is indispensable, whereas the other types are

basically monochrome. Kitaoka’s explanation of a

similar illusion is that a longer-wavelength color,

such as red, is perceived by our visual system faster

than a shorter-wavelength color, such as blue

(Kitaoka, A., 2010).

However, there are some people who cannot see

an illusion of this type, perhaps because glimpsing is

not easy for them. Therefore, we developed a system

that can move an image displayed on a PC screen,

thus enabling almost everyone to see the illusion.

3 THE SYSTEM

TO SHAKE IMAGES

We used an ordinary Windows PC to shake illusion

images, as shown in Fig. 1.

Figure 1: Our program running on a PC.

(LCD: Mitsubishi RDT 202 WM 20.1 inch, 1680 × 1050).

Our program was written in C language and was

developed with Microsoft Visual Studio 2008. The

program reads the bmp file of a still image that

causes an optical illusion and then displays the

image on the screen. The image is updated tens of

times per second and the position of the image on

the screen is changed bit by bit by controlling an

interval timer. This allows the shaken images to be

displayed.

As shown in Fig. 2, the orientation of the shake

can be designated as horizontal, vertical, diagonal

(top left to bottom right and top right to bottom left),

right rotation, and left rotation.

Figure 2: Orientation of the shake.

As shown in Fig. 3, there are two patterns of

movement: sine wave and triangular wave. The

cycle time of the wave can be varied from 10 to

2000 ms. The height of wave movement can be set

from 0 to 100 pixels.

Figure 3: The case of vertical shake.

4 EXPERIMENTS

In the experiments, we primarily used a basic Fraser-

Wilcox illusion (Fig. 4) and an optimized Fraser-

Wilcox illusion Type V (Fig. 5). The synthesis of a

pattern like the one in Fig. 4 is relatively

straightforward because it consists of simple

gradient patterns. On the other hand, a pattern like

the one in Fig. 5 is rather complex because the

variation of the color plays important roles.

Although the outline of the brightness variation

of the figure that causes this type of illusion has

been described online (Kitaoka, A., 2010), we could

not find detailed information on the variation of the

color anywhere. Therefore, the test patterns used in

our experiment were created from scratch through

trial and error by referring to published papers and

Web pages.

Horizontal Vertical Diagonal

(top left to

bottom right)

Diagonal

(top right to

bottom left)

Right

rotation

Left

rotation

VISAPP 2011 - International Conference on Computer Vision Theory and Applications

406

Figure 4: Basic Fraser-Wilcox illusion.

Figure 5: Optimized Fraser-Wilcox illusion type V.

One of the circular patterns included in Fig. 5

consists of two or more rings that share a center. If a

ring is taken out, as shown in Fig. 6, it has

periodicity along the angle θ. When one cycle is

taken out, each brightness value of the RGB

components is a function of θ because the distance

from the center does not affect the value.

A very interesting question is what kind of

functions should be used in order to obtain the

illusion. We concluded that the green component

should be at least a constant, and its desirable value

is zero. As for the red and blue components, various

functions of θ are possible. For example, sine waves

of the same frequency but different phases, one for

the red component and the other for the blue, are

acceptable, although the illusion is not very strong.

On the other hand, our experiment showed that a

relatively strong illusion can be obtained when the

functions shown in Fig. 7 are used. Please note that

the vertical axis of Fig. 7 has been normalized by the

full brightness and the horizontal axis has been

normalized by the single cycle.

Experimental results showed that the Type V

caused a stronger illusion to be felt. The results also

suggest that an appropriate viewing angle is

necessary for perceiving a strong illusion. In other

words, the viewing angle greatly influences the

effect of the enhancement. For example the rotation

could be seen when the distance between the eyes

and the display was about 30 cm. However, no

rotation could be seen when the distance was about 2

Figure 6: A ring extracted from Fig. 5.

Figure 7: An example of variation of RGB values, each of

which has the function of θ.

When the wave height is not large to some

degree, the illusion is not caused. When the

frequency of the vibration is high, the illusion grows.

However, when the pulse height is small, the illusion

is not generated. The ideal wave height varies with

individuals, but overall it is about 50 pixels. When

the pulse height grows, it becomes difficult to

determine whether the illusion is caused or not.

A supplementary experiment was conducted for

comparison. In this experiment, a hard copy of the

illusion was printed with an ordinary inkjet printer

(Canon MP960) and then hung from a clipboard

with a weight and a spring, as shown in Fig. 8. It

was then shaken by hand.

We confirmed that the optical illusion could be

felt even with this simple system. This means that

the enhancement effect of illusion described in this

paper is not limited to a PC screen. However, the

strength of the illusion with the hand shake system

was weaker than the one with the automatic shake

0.0 0.2

0.4

0.6 0.8

1.0

0.0

1.0

Normalized θ

0.5

Blue component

0.75

One cycle

Green component is

always zero.

Red component

Red Magenta Purple Red

Normalized

brightness

AUTOMATIC SHAKE TO ENHANCE FRASER-WILCOX ILLUSIONS

407

system. The reason for this is currently being

investigated. Presumed causes are a too low

frequency of the shake or the change of color.

Figure 8: Manual shake system with hard copy.

5 CONCLUSIONS

We developed a new technique for enhancing

illusory motion and were able to create a very strong

illusion by shaking a picture of the illusion that was

displayed on a PC screen by a programmed control.

A particularly strong illusion was observed in

Kitaoka's optimized Fraser-Wilcox illusion Type V.

In contrast, in the famous rotating snakes illusion

(Kitaoka, A., 2003), the enforcing effect of the shake

was barely perceived. Among the six “optimized”

Fraser-Wilcox illusions classified by Kitaoka, only

the type V showed considerable enhancement. This

result seems to suggest the possibility that the Type

V is fundamentally different from the others.

Because the Fraser-Wilcox optical illusion is

slight, experimentation has been rather difficult up

to now unless special devices are used. However, the

strength of the optical illusion can be boosted by

introducing the “automatic shake” technique

proposed in this paper, thus enabling easier

experiments. Therefore, we expect this technique to

significantly contribute to the advancement of

illusory motion research.

ACKNOWLEDGEMENTS

The authors would like to thank Dr. Masahiro

Suzuki of the Kanagawa Institute of Technology for

his valuable comments.

REFERENCES

Fraser, A. and Wilcox, K. J. (1979). Perception of illusory

movement, In Nature 281, pp. 565–566.

Faubert, J. and Herbert, A. M. (1997). The peripheral drift

illusion: A motion illusion in the visual periphery,

Perception, Volume 28, pp.108–121.

Fermüller, C., Ji, H., and Kitaoka, A. (2010). Illusory

motion due to causal time filtering. Vision Research,

Volume 50, Issue 3, pp. 315–329.

Kitaoka, A. and Ashida, H. (2003). Phenomenal

characteristics of the peripheral drift illusion, VISION,

15, pp. 261–262.

Kitaoka, A. (2003). Rotating snakes. http://www.psy.

ritsumei.ac.jp/~akitaoka/rotsnakee.html.

Kitaoka, A. (2009). Optimized Fraser-Wilcox Illusion

Type V, http://www.psy.ritsumei.ac.jp/~akitaoka/

OFWtypeV.html (in Japanese).

Kitaoka, A. (2010). A brief classification of colour

illusions, Colour: Design & Creativity, Volume 5, No.

3, pp. 1–9.

Murakami, I., Kitaoka, A., and Ashida, H. (2006). A

positive correlation between fixation instability and

the strength of illusory motion in a static display,

Vision Research, Volume 46, Issue 15, July 2006, pp.

2421–2431.

VISAPP 2011 - International Conference on Computer Vision Theory and Applications

408