Successfully reported this slideshow.

User Study of the SADIe Transcoding Engine

612 views

Published on

The World Wide Web (Web) is a visually complex, dynamic, multimedia system that can be inaccessible to people with visual impairments. SADIe addresses this problem by using Semantic Web technologies to explicate implicit visual structures through a combination of an upper and lower ontology. This is then used to apply transcoding to a range of Websites. This paper describes a user evaluation that was performed using the SADIe system. Four users were presented with a series of Web pages, some having been adapted using SADIe's transcoding functionality and others retaining in their original state. The results of the evaluation showed that providing answers to a fact based question could be achieved more quickly when the information on the page was exposed via SADIe's transcoding. The data obtained during the experiment was analysed and shown to be statistically significant. This suggests that the transcoding techniques offered by SADIe can assist visually impaired users accessing content on the Web.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

User Study of the SADIe Transcoding Engine

  1. 1. User Study of SADIe Transcoding Engine<br />Darren Lunn, Sean Bechhofer and Simon Harper<br />
  2. 2. Summary<br />SADIe transcodes content based upon meaning of the visual rendering.<br />User evaluation supports SADIe transcoding as a way to aid visually impaired users.<br />Observations from the study also provide insights into further transformations.<br />2<br />
  3. 3. The Web<br />Focus on presenting information visually.<br />Images.<br />Columns.<br />Chunks.<br />Some knowledge is only available implicitly from the page rendering.<br />3<br />
  4. 4. Implicit Knowledge<br />Advertisement<br />Banner<br />Menu<br />Main Content<br />4<br />
  5. 5. Assistive Technologies<br />Visually impaired users can make use of assistive technologies e.g. screen readers.<br />Render pages sequentially in audio.<br />Achieved by accessing the underlying HTML.<br />Focus on visual presentation rather than content hampers this.<br />Particularly if attention is not paid to coherent design.<br />Subtleties of visual presentation can be lost.<br />5<br />
  6. 6. Assistive Technologies<br />Traversal of content is serial.<br />Top-to-bottom.<br />Left-to-right.<br />Important information may not be encountered until later on.<br />Some information such as menus can be repeated for every page.<br />6<br />
  7. 7. Transcoding<br />Rules / Patterns.<br />Can adapt a large number of pages.<br />Can suffer from reduced accuracy.<br />Annotation.<br />Accurate.<br />Time consuming as every page is annotated. <br />7<br />
  8. 8. SADIe Approach<br />The visual rendering of a Web element informs the user of its purpose.<br />The CSS defines the visual rendering.<br />Identifying the purpose of the CSS definition implicitly defines the purpose of the Web element.<br />A single CSS definition is applied to every page within the Website.<br />8<br />
  9. 9. SADIe<br />Banner<br />Heading<br />Menu<br />Banner<br />Overview<br />Heading<br />Overview<br />Main Story<br />Tabs<br />Advertisement<br /> Main Story<br />Image<br />9<br />
  10. 10. Is This Useful?<br />Previously technically feasible of SADIe.<br />Conducted user studies to support transcoding from user perspective.<br />Identify future transcoding areas to investigate.<br />10<br />
  11. 11. Remote Or Lab-Based Study?<br />Remote Studies.<br />Long term.<br />Potentially more users.<br />Can see what users did.<br />Lab-based study.<br />More controlled.<br />Glean information from users.<br />Can understand why users performed actions.<br />11<br />
  12. 12. Stimuli<br />Eight text-based Websites.<br />News; Blogs; Search results.<br />Taken from Alexa top 100.<br />All pages transcoded in the same way.<br />User presented with original or modified version.<br />Tasks involved finding information on the page.<br />12<br />
  13. 13. Users<br />Four visually impaired users<br />Frequent computer users.<br />Used the Web daily.<br />All used JAWS screen reader.<br />13<br />
  14. 14. Tasks<br />Users were presented with a page.<br />They did not know which page.<br />They did not know which version.<br />Users were asked to complete a task.<br />E.g. “What is the first search result”.<br />Task completion time was calculated.<br />14<br />
  15. 15. User Interaction<br />Task index page<br />Page 3<br />15<br />
  16. 16. Time to Task Completion<br />16<br />
  17. 17. T-Test Analysis<br />Mean Times<br />Original: 163.25<br />SADIe: 54.19<br />t(7.6) = -2.14<br />p = 0.033<br />As p < 0.05 statistically significant<br />Supports the use of our transcoding<br />17<br />
  18. 18. A Good Start Point<br />Evaluation supports the use of SADIe transcoding.<br />Defluff; Reorder; Menu.<br />User discussions and observations highlight future areas of investigation.<br />18<br />
  19. 19. Importance of Structure<br />Participants frequently discussed the importance of page structure.<br />“If you told me these were the results with a heading… that would help me.”- P1 <br />“When you say heading level 4 or whatever, you need to describe what is under it.”- P3<br />19<br />
  20. 20. Importance of Structure<br />P3 found list structures much easier to navigate than paragraphs of text.<br />Scope for not only reorganising page but restructuring elements.<br />20<br />
  21. 21. Importance of Cues<br />BBC News page required users to find story headline.<br />SADIe version took longer.<br />P2 reached headline within 3 seconds.<br />Did not know it was the headline as page had changed.<br />Required cue of “Africa” before realising they had gone too far.<br />21<br />
  22. 22. Importance of Cues<br />22<br />
  23. 23. Importance of Cues<br />P1 identified search results from the word “cached.”<br />P3 knew they had gone too far from the word “feedback.”<br />Can we make use of these cues to guide people around the page?<br />23<br />
  24. 24. User Strategies<br />“Basic strategy is to use ‘find’ but this only really works when you know the page.” – P1<br />P2 quickly tabbed through the page to get an overview of the content.<br />Can these strategies be utilised to help users?<br />24<br />
  25. 25. Summary<br />SADIe transcodes content based upon meaning of the visual rendering.<br />User evaluation supports SADIe transcoding as a way to aid visually impaired users.<br />Observations from the study provide insights into further transformations.<br />25<br />
  26. 26. Questions?<br />http://hcw.cs.manchester.ac.uk/<br />darren.lunn@cs.manchester.ac.uk<br />

×