Slide 1 - IBM

bricklayerbelchedInternet και Εφαρμογές Web

5 Φεβ 2013 (πριν από 4 χρόνια και 9 μήνες)

262 εμφανίσεις

©2008 IBM Corporation



Improving the Web Browsing Environment


for Dyslexics by Elaborating Functionalities

2009
-
03
-
20

Reiko Nagatsuma, Shinji Iizuka, Masakazu Takizawa, Takuya
Ohko, Takuji Wada, Takashi Saito

IBM Accessibility Center, Tokyo Research Laboratory, Japan


©2009 IBM Corporation

2

Agenda



Motivation

-
Present status of the studies for people with Learning Disabilities
(LDs) *

-
IBM Easy Web Browsing
TM

(EWB) as a baseline tool



Improvement of “EWB for LDs”




Pilot study

-
Detail of the study

-
Results of the interviews



Conclusion





Q&A

*As you know, the meaning of LD include not only dyslexia disabilities but math, nonverbal LD and so on.

Our team will deal with the problems of LD in the future, though now we treat only the problem of dyslexia.

IBM, Easy Web Browsing are trademarks of International Business Machines Corporation in the United States, other
countries, or both.

©2009 IBM Corporation

3

Motivation and Aims

Motivation


Background



IBM Accessibility center is concerned with PwD.

-

research and develop products for easy access to Information Technology (IT).



LD: Difficulties in information processing


=> A potential area where IT could be used



Highly affected population


LD users have many difficulties in various cognitive areas



Governmental policy for developmental disorders in Japan



Recent progress in brain science


Current Status



Little research into the efficacy or sufficiency of existing IT


Challenge



Identify the potentials of assistive technology (AT)



Find out which AT is valuable for them



Find out the hidden needs for LDs thorough their opinions


=>
How about enhancements to our existing software for LD users?

©2009 IBM Corporation

4

Easy Web Browsing


a baseline tool

What is Easy Web Browsing
TM
(EWB)?

An assistive technology which provides
user
-
friendly operations and accessibility
features that enable a variety of end
-
users
to easily access websites.

There are various barriers.

Difficult to read
small font




Difficult to read

bad color
combinations


Difficult to read
long articles

Read aloud at
the location
pointed at with
the mouse

Customize the
background and
text colors

Enlarge the indicated text

Target Users: senior, Novice IT users, users who have limited
vision …

©2009 IBM Corporation

5

Extensions for LDs

How to extend EWB for LD users?

*
2) special functions for the Japanese language

*
1) Sally E. Shaywitz, R
eading
D
isorder
s, Nikkei Science,
Feb.1997.

Potential functions for LDs

Ruler

Highlight / Masking

Reading out

Magnify

Word separation (*2)

Color changing

Text to Image

Animation Suppressing

Ruby for Kanji (*2)

Line / Word / Letter spacing

Dictionary / Search

Font
-
style changing

Colored overlay

Visual System

Phoneme System

Word Recognition
System

Semantic System

-

general knowledge

-

illation

-

concept formation

-

vocabulary

Sentence Comprehension Process (*1)

[kuh aoh tuh]

cat

a furry animal
that has a long
tail and sharp
claws.

c a t

contents

EWB

1.3

EWB

2.0

EWB

LD

Enhance

Enhance

Enhance

Enhance

Enhance

Enhance

New

New

New

New

New

©2009 IBM Corporation

6

Pilot study

Pilot study for LD users

Aims

To learn what functions are valid

To verify the efficacy of this approach

Assistive
Technology

Enhanced EWB for LDs.

Enhancements depend on the reading model and the
opinions of the LD users.

Participants

1
st

test: 10 people, 2
nd

test: 9 people (mostly same users)

The study participants were all adults and all Japanese.

(Using page is written in Japanese.)

Most have LDs, and some have also ADHD.

Cooperators

NPO EDGE (Prominent Japanese organization
supporting people with LDs)

Schedule

Aug. 2008 to Dec. 2008

©2009 IBM Corporation

7

Evaluators of the pilot study

Summary of the participants

They also have writing problems like confused p and q, or “time” and
tiem

and so on.

Reading and writing problems self
-
assessed. Average age: 37.

ID

1
st

2
nd


Sex

Age

Reading Problems

A

Y

Y

M

20s

Unfocused / Cause irritation

B

Y

Y

M

30s

Repeating same line / Weak comprehension

C

Y

Y

M

30s

Blurring / Cause irritation

D

Y

Y

M

20s

Blurring / Windy sentences

E

Y

Y

F

50s

Collapsed

F

Y

Y

M

50s

Overlapping words / Cause irritation

G

Y

Y

F

30s

See double / Flow of words whirl about

H

Y

N

M

30s

Weak comprehension of words read

I

Y

Y

M

50s

Windy sentences / Overlapping words

J

Y

Y

F

20s

Slow oral reading / Cause irritation

©2009 IBM Corporation

8

Pilot study
-

Result

What functions are most needed ?

Their needs vary widely and they need a more flexible interface.

(*1) Ranking score = [number of top 1] x 8 + [number of top 2]

x 4 + [number of top 3] x 2


Each participants were asked to select top 3 functions which they wanted most.



Rank

Function

Do you need it ?


[# of persons]

Weighted
Scores

1
st

(*1)

Weighted
Scores

2
nd

(*1)

Yes

Occasionally

No

1

Read out Loud

5

3

1

38

34

2

Colored Overlay

3

5

1

-

28

3

Line Height

5

4

0

20

24

4

Ruler

3

5

1

28

10

5

Highlight / Masking

3

2

4

16

4

6

Furigana(Ruby)

3

3

3

14

2

7

Word Separation

2

4

3

12

2

8

Color Changes

2

5

2

4

0

9

Page Map

1

5

3

2

4

10

Image zoom

0

2

7

2

0

©2009 IBM Corporation

9

1: Read Out Loud Function

Summary of the participants’ comment


By far the most needed


When chosen, this function is almost always selected
as the “Best” function.


Two main comments for reading out loud


-


To understand content with a voice because of
difficulty in reading characters


-


To compare the characters on the PC with a voice
sound (to learn the words)


They have additional requirements.

One strong opinion is….

ID

Do you need it?

A

Occasionally

B

Yes

C

Yes

D

Occasionally

E

No

F

Yes

G

Yes

H

Occasionally

I

Yes

J

Yes

1


Read Out Loud

2


Colored Overlay

3


Line Height

©2009 IBM Corporation

10

A candidate of effective display method

-

KARAOKE

There was a small school near the river.

Summary of the participants’ comments


Karaoke
-
like indicator is liked by most users


Two main comments on karaoke feature:


-

Makes it easier to follow the characters with their
eyes (similar to tracing a line with their fingers)


-

Easier to understand the meaning of the
sentences. (Easier to combine the words and
corresponding sounds.)

ID

Display method

A

Karaoke

B

Karaoke

C

Karaoke

D

Karaoke

E

No need to
read aloud

F

(Other)

G

Karaoke

I

Karaoke

J

Karaoke

1


Read Out Loud

2


Colored Overlay

3


Line Height

©2009 IBM Corporation

11

2:

Colored Overlay

The Stardust spacecraft is

on a mission to collect

interstellar dust and

samples from comets.

Summary of the participants’ comment


One of the most preferred functions


It was significantly more effective than
interviewees expectations


Two main comments are:


1) Stabilizes how the characters look



irritation
decreased with lower contrast


2) Makes words clearer

-

characters are


crisper and sharper and seem to float


*Participants A&C choose 2 colors:

1
-

Japanese

2


English

ID

Do you need it?

color

A

Yes

1:yellow

2:light blue

B

No

C

Occasionally

1:red

2:light blue

D

Yes

yellow

E

Occasionally

yellow

F

Yes

yellow

G

Yes

yellow

I

Yes

yellow

J

Occasionally

yellow

1


Read Out Loud

2


Colored Overlay

3


Line Height

©2009 IBM Corporation

12

3: Line Height

Summary of the participants’ comment


Easier to follow the sentences with their eyes


Reduces confusion with neighboring lines


Easier to grasp the shapes of words


Easier to see stable word shapes


We did not enhance this parameter, but are
just reporting selected comments.

L

a

u

n

c

h

e

d

i

n

1

9

9

9

,

t

h

e

S

t

a

r

d

u

s

t

ID

Do you need it?

Ratio

A

Occasionally

200%

B

Yes

400%

C

Yes

300%

D

Occasionally

170%

E

Yes

300%

F

Yes

170%

G

Yes

500%*

G

Yes

170%

H

Occasionally

-

s

p

a

c

e

c

r

a

f

t

i

s

o

n

a

m

i

s

s

on

t

o

c

o

l

l

e

c

t

i

n

t

e

r

s

t

e

l

l

a

r

d

u

s

t

a

n

d

s

a

m

p

l

e

s

f

r

o

m

c

o

m

e

t

s

1


Read Out Loud

2


Colored Overlay

3


Line Height

©2009 IBM Corporation

13

corresponding functions

read out loud

Does the reading out loud function
substitute for phonetic processing?


The reading out loud function seems to
have a positive influence on visual
processing and word recognition.

Visible?

Meaning?

Visual System

Phoneme System

Word Recognition
System

Semantic System

-

general knowledge

-

reasoning

-

concept formation

-

vocabulary

Sentence Comprehension Process (*1)

*
1) Sally E. Shaywitz, R
eading
D
isorder
s, Nikkei Science,
Feb.1997.

[kuh aoh tuh]

cat

a furry animal
that has a long
tail and sharp
claws.

c a t

contents

eye

Adaptation of adequate model for our assistive functions.

(1) Phoneme System

Considerations
-

1

©2009 IBM Corporation

14

Adaptation of adequate model for our assistive functions.


(2) Visual system

Considerations
-

2

corresponding functions

Colored Overlay


Irlen Syndrome?

Line height

Visual System

Phoneme System

Word Recognition
System

Semantic System

-

general knowledge

-

illation

-

concept formation

-

vocabulary

Sentence Comprehension Process (*1)

*
1) Refer to : Sally E. Shaywitz, R
eading
D
isorder

,NIKKEI
SCIENCE , Feb.1997.

[kuh aoh tuh]

cat

a furry animal
that has a long
tail and sharp
claws.

c a t

contents

eye

©2009 IBM Corporation

15

Demonstration

©2009 IBM Corporation

16

Assessment of efficacy

1: Purpose:


To study the efficacy

of ATs.

2: Metrics:


Reading speed and fluency.

3: Functions:

ATs they selected.

ID

Reading speed

Improve
-
ments

Variation
(seconds)

ATs

C

7%

2.7

Colored
overlay

D

21%

7.7

Ruler

F

3%

1.0

Colored
overlay

G

20%

8.7

Magnify &
Ruby

J

11%

8.2

Colored
overlay

without AT(s)

with AT(s)

38.8

36.1

36.8

29.1

37.3

36.3

43.3

34.6

71.3

63.1

©2009 IBM Corporation

17

Conclusion

Conclusion


We improved some of the functions in IBM Easy Web Browsing (EWB),
which was developed for people with limited vision and senior citizens,
to help LD users.


We evaluated the efficacy of the functions.


Our conclusions:

-

ATs are certainly helpful, though their needs vary widely.

-

The top 3 functions were:



1: Reading out loud, 2: Colored overlays (yellow), and 3: Line height

-

Their reading speed was better with the ATs

The users wanted more flexible customizations and combinations of the ATs.


©2009 IBM Corporation

18

Conclusion

To next stage…


Why are certain ATs effective? How are brains working?



The functions still need improvements.


We want a highly customizable UI programming framework.


User modeling would support fast and easy individual adaptations.

©2009 IBM Corporation

19

The Most Impressive Comment

The ease of access level for Web browsing is

directly connected to the quality of life.


Therefore this factor influences one's thinking of

whether to become pessimistic or optimistic for

the inevitable stage of aging.

Mail from 6 Sep. 2008

©2009 IBM Corporation

20

Q & A

©2009 IBM Corporation

21

Thank You Very Much!

contact:
reiko@jp.ibm.com


20
,

Mar, 15:30
-

17:00, demonstration of EWB for LD at #270


©2009 IBM Corporation

22

Other functions

ruler

Underline. It helps to follow the
line with their eyes.

Magnifi

-
cation

Line space is wider with
magnification, but the recognition
of the characters becomes
difficult.

Furigana

(Japanese

Only)

Furigana will help to read
however improvements are
needed for allocation and
coloration.

Word
separating

(Japanese

Only)

Word separating allows easy
word recognition by separating
Japanese characters.

Colour
Change

Colour change has the same
effect as coloured overlay








さん















岸に小さな学校がありました。

岸に

小さな


学校が


ありました

There was a little school near the river

There was a little school near the river

Since discovering her own daughter's
struggle with reading and Irlen Syndrome


Since discovering her own daughter's
struggle with reading and Irlen Syndrome


Since discovering her own
daughter's struggle with reading
and Irlen Syndrome

©2009 IBM Corporation

23

Evaluators of the pilot study

Corresponding table between reading problems and AT

ID

Reading Problems

Read Out
Loud

KARA

-
OKE

Colored
Overlay

Line
Height

A

Unfocused / Cause irritation

Occasionally

Yes

Yes

Occasionally

B

Repeating same line /

Weak comprehension

Yes

Yes

No

Yes

C

Blurring / Cause irritation

Yes

Yes

Occasionally

Yes

D

Blurring / Windy sentences

Occasionally

Yes

Yes

Occasionally

E

Collapsed characters

No

No

Occasionally

Yes

F

Overlapping words

Cause irritation

Yes

Yes

Yes

Yes

G

See double

Flow of words whirl about

Yes

Yes

Yes

Yes

H

Weak comprehension

Occasionally

-

-

-

I

Windy sentences / Overlapping
words

Yes

Yes

Yes

Yes

J

Slow oral reading

Cause irritation

Yes

Occasionally

Occasionally

Occasionally

©2009 IBM Corporation

24

1: Purpose:


To study the efficacy

of ATs.

2: Metrics:


Reading speed and fluency.

3: Functions:

ATs they selected.

ID

Reading speed

Improve
-
ments

Variation
(seconds)

ATs

A

25%

14.2

Colored
overlay

B

24%

21.0

Line Height

E

19%

6.5

Line Height

without AT(s)

with AT(s)

56.4

42.2

86.0

65.0

34.6

28.1

©2009 IBM Corporation

25

Assessment of efficacy

Assessment pages