and everything else Richard E Ladner and Jeffrey

and everything else? Richard E. Ladner and Jeffrey P. Bigham Work with Ryan Kaminsky, Gordon Hempton, Oscar Danielsson University of Washington Computer Science & Engineering

Web Accessibility Overview Accessibility Affects n People who are blind n People with visual impairments n People who are Deaf or hard of hearing n People with learning disabilities n People who are physically impaired 2

Web Accessibility Overview Accessibility Affects (cont. ) n People who use cell phones n People who use text browsers n Information extraction 3

Web Accessibility Overview Standards for Developers n W 3 C Web Content Accessibility Guidelines n Section 508 of the U. S. Rehabilitation Act n Americans with Diabilities Act (ADA) 4

Web Accessibility Overview Accessible Browsing n Screen readers, refreshable Braille displays Consider Linear Display n Separate presentation from meaning n No vision or mouse required n Visual content requires an alternative n 5

Web Accessibility Overview Images n Images cannot be read directly n W 3 C accessibility standard ¨ “Provide n a text equivalent for every non-text element” What if no alternative text? ¨ Nothing ¨ Filename (060315_banner_253 x 100. gif) ¨ Link address (www. cs. washington. edu or /subdir/) 6

7

Update Address /olc/pub/YALE/oldintro. cgi 8

Cornell CS Webpage 9

Making Images Accessible Part II: Accessible Images n Web Studies n Providing Labels n Web. In. Sight System n Evaluation n Developers 10

Making Images Accessible Web Studies: All Images != n Significant images need alternative text ¨ alt, title, and longdesc HTML attributes n Insignificant images need empty alt text ¨ Decorative or structural alt=“”> height=“ 1” 11

Making Images Accessible Image Significance More than one color and both dimensions > 10 pixels n An associated action (clickable, etc. ) n 12

Making Images Accessible Web Studies n Previous studies ¨ All images: n ¨ Significant images: n n 27. 9%[1], 47. 7%[2], and 49. 4%[2] 76. 9%[3] Concerns Variation ¨ Consideration of Image Significance and Popularity ¨ [1] T. C. Craven. “Some features of alt text associated with images in web pages. ” (Information Research, Volume 11, 2006). [2] Luis von Ahn et al. “Improving accessibility of the web with a computer game. ” (CHI 2006) [3] Helen Petrie et al. “Describing images on the web: a survey of current practice and prospects for 13 the future. ” (HCII 2005)

Making Images Accessible Web Site Study Group Significant Pages > 90% Pages Images High-traffic 39. 6% 21. 8% 500 32913 Computer Science 52. 5% 27. 0% 158 4233 Universities 61. 5% 51. 5% 100 3910 U. S. Federal Agencies 74. 8% 55. 9% 137 5902 U. S. States 82. 5% 52. 9% 51 2707 Percentage of significant images provided alternative text, pages with over 90% of significant images provided alternative text, number of web sites in group, 14 and number of images examined.

Making Images Accessible Web Traffic Study n University of Washington CSE Department Traffic ¨ ~1 week ¨ 11, 989, 898 images including duplicates ¨ 40. 8% significant ¨ 63. 2% alt text Significant images with alternative text. Significant images without alternative text. 15

Part II: Accessible Images n Web Studies n Providing Labels n Web. In. Sight System n Evaluation n Developers 16

Making Images Accessible Providing Labels: Context Labeling n Many important images are links Linked page often describes image ¨ What happens if you click ¨ src=“p 523. gif” alt=“People of UW”> People … 17

Making Images Accessible Providing Labels: OCR Labeling (Optical Character Recognition) Improvement through Color Clustering[4] Color New Image Text Produced , , n Improves recognition 25% relative to base OCR! Register now! [4] Jain et al. “Automatic text location in images and video frames. ” (ICPR 1998) 18

Making Images Accessible Providing Labels: Human Labeling [5] n n [6] Humans are best Recent games compel accurate labeling Web. In. Sight database has only 10, 000 images Could do this on demand [5] Ahn et al. “Labeling images with a computer game. ” (CHI 2004) [6] Ahn et al. “Improving the accessibility of the web with a computer game. ” (CHI 2006) 19

20

Making Images Accessible Part II: Accessible Images n Web Studies n Providing Labels n Web. In. Sight System n Evaluation n Developers 21

Making Images Accessible Web. In. Sight System n Tasks Coordinate multiple labeling sources ¨ Insert alternative text into web pages ¨ Add code to insert alternative text later ¨ n Features Browsing speed preserved ¨ Alternative text available when formulated ¨ Immediate availability next time ¨ 22

Making Images Accessible The Interne t Context Labeling OCR Labeling Proxy Human Labeling Database Blind User 23

Making Images Accessible Labeling Service OCR Labeling The Interne t Human Labeling Database Extension Blind User Context Labeling 24

Making Images Accessible Concerns n Accuracy n Distribution of Tasks – who does what? n Authorization – who can use the system? n Privacy n Copyright 25

Part II: Accessible Images n Web Studies n Providing Labels n Web. In. Sight System n Evaluation n Developers 26

Making Images Accessible Evaluation n Measuring System Performance ¨ ¨ ¨ n Web. In. Sight tested on web pages from web site study Used Context and OCR Labelers Labeled 43. 2% of unlabeled, significant images Sampled 2500 for manual evaluation 94. 1% were correct Proper Precision/Recall Trade-off 27

Making Images Accessible 28

Part II: Accessible Images n Web Studies n Providing Labels n Web. In. Sight System n Evaluation n Developers 29

Making Images Accessible Developers: Prior Work n A-Prompt U of Toronto as part of W 3 C initiative, 1999 ¨ Registry for alternative text ¨ Provides suggestions using heuristics on filenames ¨ n ALTifier Proxy-based system ¨ Used filename/URL as alt text ¨ 30

Web. In. Sight Developer Video 31

Making Images Accessible Conclusion n Lack of alternative text is pervasive n Web. In. Sight formulates & inserts alt. text n Appropriate precision/recall tradeoff n Users and developers can use same technology 32

Future Research Part III: Future Research n Support Web Users and Developers n Automation and Suggestions n Independence n Sharing and Collaboration 33

Future Research Understanding our users n Blind web users Remote observation with proxy server ¨ User diaries ¨ n Web developers Focus groups ¨ Surveys ¨ 34

Future Research Technical Challenges n Relaying Content Structure ¨ n Dynamic Content ¨ n DHTML, mouse overs Rich Internet Applications/Web Applications ¨ n tables, div, columns e-mail, word processing, spreadsheets Requires new ways of reading the web 35

36

Future Research Scripting Accessibility Greasemonkey reshapes the web n Accessmonkey facilitates accessibility n Getting technology to people ¨ Multiple platforms and implementations ¨ A conduit for collaboration ¨ Web users and developers share technology ¨ 37

Future Research Independence n Automation means independence n Helping users create scripts n Helping users share scripts 38

Related Projects Part IV: Related Projects 39

Tactile Graphics Graphic Translation text extract preprocess original scanned image location file clean image 16 100. 000000 1. 923077 1. 953125 121 45 140 69 0 3. 141593 pure graphic text image 40

Tactile Graphics Graphic Translation location file pure graphic text image 16 100. 000000 1. 923077 1. 953125 121 45 140 69 0 3. 141593 text y (0, 20) x=15 15 10 5 O x 5 10 15 20 20 x+y=20 (15, 0) (15, 5) Braille y (#0, #20) x. k#15 #10 #5 O x #5 #10 #15 #20 x+y. k#20 (#15, #0) (#15, #5) 41

Mobile. ASL Project n ASL communication using video cell phones over current U. S. cell phone network Challenges: n n Limited network bandwidth Limited processing power on cell phones 42

Web. In. Sight http: //webinsight. cs. washington. edu Thanks to: Luis von Ahn, Scott Rose, Steve Gribble and NSF. 43