The project is aimed at developing a natural sounding TTS engine for Indian languages. The project involved developing various spoken language processing modules like Font Converters, Text Normalization, POS Tagger for Prosody pause prediction, grapheme-phoneme converter, Stemmer, optimal text and unit selection etc. Under the framework of data driven speech synthesis, wherein speech is synthesized by concatenating optimal units selected from a large corpus of recorded utterances. This approach produces very natural sounding speech.
Role Played: Building TTS for Hindi, Telugu, Tamil, Bengali and Malayalam. Here I am working with generic frameworks like Festival and Festvox and I have created my own modules for language processing, using C++ and PERL.
[Project Instructor :: Mr. Kishore Prahallad]
[Project Members :: Venkatesh Keri, Anand Arokia Raj, E Veera Raghavendra, Sathish Chandra Pammi, Tanuja Sarkar, Srinivas Desai, Tripuramallu Bhaskar, Veenadhiri Pammi ]

Question Answering is a specialized form of information retrieval. Given a collection of documents, a Question Answering system attempts to retrieve correct answers to questions posed in natural language. Factoid based question answering requires question answering systems to be able to answer questions in a single fact. For example “Where is the TajMahal?", “How tall is the Eiffel Tower?”, and “Where is Microsoft based?” are all examples of factoid questions. This Project involves developing various modules like, Document Classification, Document Indexing and Retrieval, Keywords Extraction, Question Analysis, Answer Ranking and Answer Extraction.
Role Played: Here I am working with generic frameworks like Lucene, Mini-Parser, IIIT NER and machine learning algorithms CART, Language Modeling and I have created my own systems using C++, AJAX and PERL.
[Project Instructor :: Dr. Vasudev Varma]
[Project Members :: Praneeth M Shistla]

Role Played: Building Acoustic Model, Language Model, Natural Language generation and Dialog manager.

Spoken dialogue systems enable users to interact with computer systems via natural and intelligent dialogues, as they would with human agents. The project involved developing various spoken language processing modules like Language modeling, Natural language understanding (keyword extraction and processing), Dialog manager (Finite automata and database query), Text normalization and Natural language generation.
Role Played: Building Language Lexical, Language Modeling and Dialog manager for Tamil and Telugu.
[Project Instructor :: Mr. Kishore Prahallad]
[Project Members :: Venkatesh Keri, Karthick]

The scope of the application is to parsing the input text data into group of phones this is used in an internal part of Text to speech system for unit selection. Most Indian languages use a syllabic script that is largely phonetic in nature thus, in most case; there is a one-to-one mapping between the grapheme (minimal text unit) and its corresponding phone. However, there is another possibility to get one-to-many mappings. Thus, a G2P converter for Indian languages needs to provide not only the basic mapping between the graphemes and the phonemes but also the rules that enhance or change these mappings.
Role Played: Building Letter to sound rules for Tamil, Hindi and Bengali using machine learning algorithms such as CART and feature extraction modules using PERL & C++.
[Project Instructor :: Mr. Kishore Prahallad]
[Project Members :: Tanuja Sarkar, Sathish Chandra Pammi]

[Project Instructor :: Mr. Kishore Prahallad]
[Project Members :: Venkatesh Keri, Sathish Chandra Pammi]

[Project Instructor :: Mr. Kishore Prahallad]
[Project Members :: Sathish Chandra Pammi]

[Project Instructor :: Mr. Kishore Prahallad]
[Project Members :: Venkatesh Keri, Sathish Chandra Pammi]

Voice Enabled Mail Reader!

---

Last Modified: December 28, 2006 1:01 PM IST