バイオロジカルモーションのスリット視(1)(第23回大会 優秀発表賞抄録)

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

ThejdPanesefournatofIls),chononticSciance

2005, VoL 24,N{].1,133-134

Summary

ofAwarded

Presentation2P46

Theperception

of

a

Point

Light

Walker

across

a

-The

perception

of

the

"human

likeness'L

slit

ChihiroAsANol

and

Yoshihisa

Rihle.voUbeiversitJJ'

OSADA

The aim of the _present study was to elarify whcther thependulum motion ol a Peint-Light

Walker

{PLW)

created a

_perception

of

it's

"human

likeness".

To _quantify the amount of

spatio-tempora]

information

rcquircd tocreate an

impression

of "human likeness",thePLW was _passed

behind a slit which varied inwidth and

the

expo$ure time was also varied.

Observers

asscssed the

"human _likeness"

of each PLW

display,

The

experimental results revealed thata distinct

impres-sion of "human

likeness"iscreated when observcrs sec more than a halfcycle of the _pendu]um

motion

from

the

PLW

for

rnore than 41O ms. These findingssuggest that _{the pendulurn} motion of aPLW conveys information which

is

relevant

to

_{theperception} of "human

]jkeness"

under certain spatjal and temporal condjtions.

Key words: PointLight Walker, slit, human likenes$,_pendulum motion

Introduction

Even

without a complete

image

ofa human

wa]k-er,

bright

dots

representing thc main

joints

of the

human body when seen against a dark background

can give us

the

_perception of one.

Johansson

(1976}

has explained the mechanisms forthis _phenomenon

within

the

frarnework of "figure"

and '`ground"

seg-mentation, This conceptual framework separates

human

walking

into

translationand pendulum

rno-tions.Thependulum motion of a

PLW

is

considered

to

be

especially

important

for

identifying

the

move-ment as human

(Owens

et aL 1994;

Shiffrar

et al.

1997).To date itremains unknown as towhether the

pendulum motion of the "`figure"

alone inherentlyhas

information which

is

suMcient to

lead

tothe

percep-tionof "human _likeness",

or not.

The

purpose of the present study was to

investi-gate:

{1}

whether the pendulum motion of a

PLW

contained

information

relevant tothe_perceptionef a

human

likeness;and

(2)

how much spatio-temporal

information was required tocreate an impression of

"human _{likeness"from the}

pendulum motion.

Methods

Experiment

1:

Spatial

conditions

Observers Fifteen observers

(4

rnen and 11

* _{Department of}

_Psychology,

_Rikkyo

_University,

34-1

Nishiikebukuro,

Toshima-ku,

Tokyo

O021

women) with normal visual acuity participated.

Stimuti Three motion _patterns of the PLW

(up-right PLW, inverted PLW, scrambled PLW) were

shown behind a slit, moving from the leftto

the

righL The scrambled _pattern,which

did

not contain

any inforrnationrelevant tothe perception of

"hu-man likeness"was constructed

by

randomly

chang-ing

the

frame

order of thePLW. Seven differentslit

widths with visual angles of O.80,1.60,2,40,3,20,4.0e,

and

4.80,

were used to limitthe spatial information

from themotion _pattern, A slit width of 22.50 corre-sponded toa "no-slit"

condition,

Procedure The observers viewed three motion

patterns which were

behind

a slitand then rated

their

impression

of "human

likeness"

of each. Thc

rating scale

for

each rnotion pattern was from grade

1

to

5,

with

grade

1indicatingthat

"human

likeness"

was not _perceivcd inthe metion _patternand _grade

5

indicating

that

"human

]ikeness" was clearly

per-ceived in the motion _pattern.

Five

trialsol the

upright PLW,

5

trialsof

the

inverted

PLW,

and

10

trials of the scrambled condition were _presented in

random order

for

each observer. Sothat the $patial

information

frorn

the

PLW

could be _quantified we

measured the

distance,

along a horizontal axis in

each frame, from the head to the rightankle. The

pendulum-like motion of _the right ankle was

repre-sented on a 2-dimensional _plane by _plotting each

ankle distance

ly-axis)

again$t thetirne

(x-axis)

which

was cnlculated frorn_the

frame

rate.

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

134 The

Japanese

Journal

of Psychonornic Science Vol.24, No. 1

5

z4e.-<E3:8

2 1o.s

Figure l. The

ness with eac

D

1.6 2.4 3.2 4.0 4.S 22S

(NoSlit) SLTTWIDTH(degree)

estimation of a

human

h

orientation condition

(Exp,

1)

Experiment

2: Temporal conditions

Observers

Six

observers

{3

men and 3 wornen)

with normal visual acuity _{participated.}

Stimuli

UprightandinvertedPLWswercusedin

a similar manner tothatof

Experiment

1.

A

control

PLW, without any

information

of a

human

likeness,

was constructed

by

moving 11 dots in a parallel

arrangement towards the right, The three motion

patterns were _passed behind a slit,frornthe leftto

the

right, with differentframe rates

(15

frames _per

second,

30

frames

_per

second, and

60

frames

_per

second) so that the exposure

time

of

the

patterns

could

be

changed.

Procedure

Five trialsof each motion pattern

were carried out ina random order and the observers rated theirimpre$sion of "human

likeness"in the

same manner as

in

Experiment

1,

Results

and

Discussion

Experiment

1: Spatial conditions

Figure 1 illustratesfor each motion pattern the

relationship between the estimation of "human

like-ness" and theslit width.

The

evaluation of

"human

likeness"increased as the s]it width was expanded

with

the

upright and

inverted

PLWs, whereas it

re-mained

low

in

_the scrambled PLW conditions. The

rating

level

reached a plateau with an upright PLW

and a slitwidth of 3.2degrees or rnore. Under these

conditions more than halfa cycle of the _pendulum

motton was observed.

Experiment2: Temporalconditions

Figure 2 indicates

how

the evaluation of "human

likeness" varicd with expansion of the slit width

under different

frame

rate conditions,

The

evalua-tionof "human

likeness"

increasedsignificantly with

5

g4

:

E3

8

2

1 (NoSlit) SLITWIDTH(degree)

Figure2.

The

estimation of a human

ness with an upright PLW, under three

conditions ef frame rate

(Exp.

2)

expansion of the slit at allframe rates. A significant

difference

was not observed

between

a rate of l5

frames

and 30

frames

_per second.

Within

the slit

range from

2,4

to

4.8

degrees

_the evaluations forthe

60 frames _per second rate were signjficantly lower

than those

for

theother

frame

rates. The mechanism

forthis

disparity

could be explained by a difference m exposure time tothe

_pendulum

motion,

When

the exposure _time was

less

than 410 rns, evaluation of "human

likeness"remained low, even when the slit

width was 3.2degrees or more and

therc

was more

than one half-cycleof _pendulurn motion.

General

Discussion

The present study revealed that the _pendulum

motion of a

PLW

could provide information which

is

related to"human _{likeness".We found}

thatinorder

to create a distinctimpression of 'Chuman

likeness",

observers need tosee more than one

half-cycle

of the

pend ulurn motion over 410 ms,

Thus,

the _penduluin

motion of a PLW can create a clear jmpression of

"human

likeness"within aspecific spatial and

tempo-ral window.

References

Johansson,

G. 1976

Spatio-temporal

differentiation

and integration

in

visual rnotion perception: An

expenmental and theereticalanalysis of

likefunctions

in

visual

data

processing.

Rsychologi-cat

Research,

38(4L

379-393.

0wens,

D.

A.

AntonofL

R,

J.,

& Francis,E.L,1994

Biological

motion and nighttime Pedestrian

spicuity. Hbeman Factors,

36(4),

718-732.

Shiffrar,

M. Lichtey,C. &

Sheba

H.

1997

The

tionof biologicalmotion across apertures.