Multliligual Self-Voicing Browser Model for Ethiopc Scripts
No Thumbnail Available
Date
2014-04-17
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
This master’s thesis describes the design of multilingual self-voicing browser model for Ethiopic scripts. We show that efficient multilingual self-voicing browser model can be constructed for Ethiopic scripts by a purification, parsing, transcoding, and text-to-speech techniques and approaches. Self-voicing browser model research have been an active research for English and other non-Ethiopic languages for quite long time. Though, there has not been any research work done before to design and implement self-voicing browser for websites equipped with Ethiopic scripts. There are various approaches and techniques proposed for English and other languages in designing self-voicing browser. We have adopted the approach introduced by the research work of WebAnywhere for English language and extended the approach to be a multilingual. Multilingual self-voicing browser model is a web based web browsing application that allows visually impaired people access the web from anywhere. The model designed does not require any additional software to be installed on the web browsing machine other than sound card and web enabled machine. In addition, the model is not designed for specific language rather to support multiple language bases. The webpage coded in none accordance with W3C compliance was found to have crucial effect in the performance of our model. Evaluation of the model showed that webpages purified for W3C compliance were not accurately purified to exhaustively parse and generate the required content. We have learned that the webpage purification accuracy has a direct impact on the parser and text transcoder to traverse and generate all the required information. The evaluation of our model, being the first multilingual self-voicing browser model for Ethiopic scripts, has shown promising performance. The HTML purifier library improved the webpage W3C coding compliance of standard by 44% precision accuracy. The parser performance in traversing and generating DOM tree is 96% precision accuracy for webpages their HTML coding compliance improved by HTML purifier. The text transcoder achieves a precision accuracy of 100% by transcoding all parser generated DOM tree. Though, this doesn’t mean that all the required content in the webpage has been successfully transcoded as the HTML purifier only cleaned the webpage at accuracy of 44% which we do not able to get 56% of the webpage content that needs to be transcoded. In general, the approaches and libraries, we have used to design our model shows a meaningful performance considering an initial research work.
KEYWORDS: Self-voicing browser, HTML Purifier, Parser, text transcoder, Text to Speech, Website Accessibility, Website Usability
Description
Keywords
Self-Voicing Browser; HTML Purifier ;Parser; Text Transcoder; Text to Speech; Website Accessibility; Website Usability