Upload
kory-waters
View
213
Download
0
Embed Size (px)
Citation preview
Reading Aid for Visually Impaired
Veera Raghavendra, Anand Arokia Raj, Alan W Black, Kishore Prahallad, Rajeev Sangal
Language Technologies Research Center,International Institute of Information Technology, Hyderabad
www.iiit.ac.in
Agenda
• Motivation• Challenges• Architecture of RAVI• Sub Systems• What is working and progressing• Salient features of RAVI
Motivation
• Visually impaired should be able to access and interact with the computer as a normal person.
• Visually impaired should be able to access computers in their native language - Indian languages
• Multi-Lingual screen reader
Challenges• Tracing the system context information • Identifying the active application and extracting the
text content • Identifying the language or script notation • Language processing or script/font conversion • Text normalization • Invoking the right TTS system • Building a high quality TTS system for a language • Real-time response of the system
WordPad Microsoft Word
Microsoft Excel
Microsoft PowerPoint
Internet Explorer
Language Based TTS Selector
Font Converters, Text Normalization Module and LTS
Text, Font Name Text, Font Name
Text
Speech Output
Text, Font Name
Phonetic Notation of Input Text
Text to Speech (TTS) Engine
Accessibility proxy & Stub
System Wide Accessibility Event
TracerKeyboard Event Tracer
Microsoft Outlook
Architecture of RAVI
Sub Systems
• Text Information Retrieval• System wide• Application specific
• Speech Synthesis• Font Conversion• Text Normalization• Letter to Sound (LTS) Rules• TTS for Indian and English Language (s)
Text Information Retrieval
• System Information Retrieval• Microsoft’s Accessibility Technology is used to retrieve the system
information• It allows application to expose the type, name, location, and current
state of all objects• And notifies any Windows event that leads to a user interface
change.
• Application specific Information Retrieval• Identifying the Active Application on the desktop• Retrieving the Text information from the Active Application
• Microsoft Active Accessibility (MSAA)• Component Object Model / ActiveX Controls • Windows Message processing
Speech Synthesis• Font Conversion
• True Type Fonts• Uses 1-256 ASCII characters to represent characters
• Character representation is different from one font to other [even in the same language]
• Separate converter required for each font
• Proprietary fonts
• Unicode
• A universal character set
• provides a unique number for each character in a language
• Supports all platforms
• Supports all the languages..Continued
• Text Normalization• This is a Taj Mahal• Welcome to Dr.A.P.J Abdul Kalam• Cost of this pen is Rs.50.• Monday 14 January 2008, 19:48 GMT• I need P3 2.6Ghz, 2GB RAM, 300 GB SATA with 17” LCD
screen.
• Letter to Sound Rules• Written form of the word is different from spoken form in some
languages like Hindi and Tamil. • Eg: /r aa m a/ is spoken as /r aa m/• /aa ch I y aa/ cab be spoken as /aa s I y aa/
..continued
• TTS for Indian and English Language (s)• Indian Languages
• A character in Indian language scripts is close to a syllable • a syllable and can be represented as C*VC* where C is a consonant
and V is a Vowel• Syllable based synthesizer has been built for Indian Languages• Collect few thousand of sentences which are phonetically balanced• Record those sentences using native and well known speaker of the
language• Create syllable database from the recorded sentences• To reduce the database between 20-40 MB, select the syllable units
which are neutral
• English• Microsoft SAPI 5.0 (default for Windows XP)
What is working and progressing• Working
• System Information Retrieval• Application specific Information Retrieval
• Font Conversion• True Type Fonts• UNICODE
• Text Normalization• Letter to Sound Rules• Indian Language TTS for Telugu
• Progressing• Indian Language TTS for Hindi, Tamil, and Kannada
Salient Features of RAVI
• Voice guided installation support • Automatically starting at system boot up and option to remove start up • Character echoing while typing • Detailed help documentation for usage and keystrokes combinations • This software integrates itself with almost all the variants of Windows
operating system • This system identifies the language, based on the current context (font
name) and automatically switches to the respective speech engine • For producing English voice, we have used “Microsoft TTS” system and
for Indian Language voices, we have used ‘Indian Language TTS’ system developed at IIIT-Hyderabad
• This software supports applications like MS Word, Notepad, WordPad, PowerPoint, Outlook and Internet Explorer
Thanks
Questions?