+ All Categories
Home > Science > B#: Chord-based Correction for Multitouch Braille Input

B#: Chord-based Correction for Multitouch Braille Input

Date post: 24-May-2015
Category:
Upload: hugo-nicolau
View: 202 times
Download: 15 times
Share this document with a friend
Popular Tags:
30
B# :: Chord-based Correction for Multitouch Braille Input HUGO NICOLAU KYLE MONTAGUE TIAGO GUERREIRO JOÃO GUERREIRO VICKI L. HANSON
Transcript
Page 1: B#: Chord-based Correction for Multitouch Braille Input

B# :: Chord-based Correction for Multitouch Braille InputHUGO NICOLAUKYLE MONTAGUETIAGO GUERREIROJOÃO GUERREIROVICKI L. HANSON

Page 2: B#: Chord-based Correction for Multitouch Braille Input

motivation :: non-visual input

Page 3: B#: Chord-based Correction for Multitouch Braille Input

motivation :: braille input

[Southern et al. 2012]

[Oliveira et al. 2011] [Azenkot et al. 2012]

Page 4: B#: Chord-based Correction for Multitouch Braille Input

motivation :: fast typing (4x)

Page 5: B#: Chord-based Correction for Multitouch Braille Input

problem :: error prone

50%

Page 6: B#: Chord-based Correction for Multitouch Braille Input

challenge :: accurate input

Page 7: B#: Chord-based Correction for Multitouch Braille Input

approach :: spellchecker

? ? ? ? ? ? ? ? ? ?

> e d k l o w u h k d _

Page 8: B#: Chord-based Correction for Multitouch Braille Input

approach :: character-level

e x i l e d u n k s

> e d k l o w u h k d _

Page 9: B#: Chord-based Correction for Multitouch Braille Input

B# :: chord-level spellchecker

h e l l o w o r l d

> e d k l o w u h k d _

Page 10: B#: Chord-based Correction for Multitouch Braille Input

B# :: data collection

Page 11: B#: Chord-based Correction for Multitouch Braille Input

B# :: data collection

11 blind participants :: ages 22 – 62 ::

3 females :: grade 1 braille ::

no touchscreen experience

Page 12: B#: Chord-based Correction for Multitouch Braille Input

major results :: substitutions

Insertions Substitutions Omissions

23%

72%

5%

Overall Error Rate: 18%

Page 13: B#: Chord-based Correction for Multitouch Braille Input

major results :: 43% single-dot error

omission

insertion

transposition

q g

i f

b k

Page 14: B#: Chord-based Correction for Multitouch Braille Input

major results :: invalid characters

! t

Page 15: B#: Chord-based Correction for Multitouch Braille Input

B# :: braille distance

[100010]

[100010]

= ?

-

Page 16: B#: Chord-based Correction for Multitouch Braille Input

distance :: damerau-levenshtein

[100] – [110] + [010] – [010]

= 1 + 0 = 1

Page 17: B#: Chord-based Correction for Multitouch Braille Input

B# :: scoring function

scoreword = wMSD(Chtyped, Chword) + ƒword

Page 18: B#: Chord-based Correction for Multitouch Braille Input

evaluation :: B#, AOSP

Page 19: B#: Chord-based Correction for Multitouch Braille Input

evaluation :: dataset

751 words :: 49% incorrect entries ::

word list with 213,133 entries

Page 20: B#: Chord-based Correction for Multitouch Braille Input

results :: B# is more accurate

1 2 3 4 5 6 7 8 9 1025%

35%

45%

55%

65%

75%

56%

75%

28%39%

Acc

ura

cy

B#

Android

Page 21: B#: Chord-based Correction for Multitouch Braille Input

results :: word correction

204 of 364

100 of 364

Page 22: B#: Chord-based Correction for Multitouch Braille Input

results :: no gain beyond 5

1 2 3 4 5 6 7 8 9 1025%

35%

45%

55%

65%

75%A

ccura

cy

B#

Android

6%

9%

Page 23: B#: Chord-based Correction for Multitouch Braille Input

results :: incorrectly corrected

Android B#

1.4%1.7%

Page 24: B#: Chord-based Correction for Multitouch Braille Input

conclusion :: B#

Page 25: B#: Chord-based Correction for Multitouch Braille Input

conclusion :: higher accuracy

2X

Page 26: B#: Chord-based Correction for Multitouch Braille Input

future :: non-visual interface

Page 27: B#: Chord-based Correction for Multitouch Braille Input

The End.HUGO [email protected], slides, and video @ http://web.ist.utl.pt/hugo.nicolau

Page 28: B#: Chord-based Correction for Multitouch Braille Input

non-visual correction :: challenges

Automatic correction?

Word-completion?

How to provide corrections?

When to provide them?

How to users explore suggestions?

Page 29: B#: Chord-based Correction for Multitouch Braille Input

B#: future improvements

Bi-grams, Tri-grams, N-grams

“Machine Learned” Distance Function

Personalized spellchecker

Leverage touch information

Page 30: B#: Chord-based Correction for Multitouch Braille Input

B#: android’s proximity matrix

Fairly complex spellchecker

Dozen of features

Worthwhile doing


Recommended