Speech and Text Understanding (STU) lab conducts research and development in Thai spoken and written language technologies. The lab is engaged in wide range of R&D activities such as natural language processing, speech recognition and synthesis, information retrieval and extraction and knowledge discovered from texts. Such activities help enable computers to interact and communicate with humans with natural language capabilities.
สารบัญ
Vision
To become the leading R&D lab in Thai spoken and written language technologies
Mission
- To achieve excellence in basic and applied research in Thai spoken and written language technologies
- To collaborate with domestic and international research organizations.
- To create and setup ecosystem for research in Thailand
- To create the greatest economics and social impact through technology transfer
Core Technology
- Speech Technology
- Signal and Speech Processing
- Speech Recognition
- Speech Synthesis
- Audio-visual technology
- Text Processing and Mining Technology
- Word Segmentation
- Information Retrieval and Search Engine
- Text Mining
Renown Project
1. VAJA: Thai Speech Synthesis
Vaja is a Thai text-to-speech software and has been extensively researched and developed since 1997. The most recent version Vaja 8.0 comes with the following ;
- Able to synthesis all Thai words even some words are not occurred in the dictionary
- Adding words to a custom dictionary such as, a named entity, etc.
- Two languages (Thai and English) with one voice
- Support multiple platform such as, Windows (Microsoft Speech APIs), Linux Enterprise (Web APIs) และ Android (Android APIs)
2. Partii: Automatic Thai Speech Recognition
Partii is a Thai speech recognition system, which converts speech signal to text. This project has been researched since 2013. The current version is Partii 4.0 comes with the following ;
- Accuracy: 80% in quiet environment and 70% in quasi-noisy environment.
- Real-time-factor: 1.5x
- Supporting client and server streaming through gRPC protocol
- Customizable system
3. S-sense: Social Media Monitoring
S-Sense is a tool for analyzing intention and sentiment on Thai social media. S-Sense can classify social media content into four intentions: request, question, announcement, and sentiment. The sentiment text can be further classified into either positive or negative. S-Sense can also analyze comparative sentence and identity association between sentiment and feature of the products or services.
- Compatible multi-file format: plain Text (txt), Microsoft Word Document (doc, docx), Portable Document Format (pdf)
- Support Thai and English document
- Detect some modified technique such as, deletion, addition, insertion, swap, etc.
- Customisable database
- Highlight suspicious sentences with similarity percentage
- Statistical report
Members & Expertise
- Vataya Chunwijitra, Ph.D (Automatic Speech Recognition, HMM-based Speech Synthesis)
- Chaianun Damrongrat, Ph.D (Information Retrieval, Computer Vision, Augmented Reality, Knowledge Management, Mobile Application, Location-Based technology)
- Ausdang Thangthai, Ph.D (Speech processing, Thai Text processing, Audiovisual Speech Synthesis)
- Kwanchiva Thangthai, Ph.D (Multi-Modal Speech Recognition, Statistical Language Modeling)
- Surasak Boonkla, Ph.D (Speech Processing, Voice Biometrics)
- Pornpimon Palingoon (Linguistics, Corpus Linguistics, Lexicography)
- Sumonmas Thatphithakkul (Linguistics and Phonetics)
- Sarawoot Kongyoung (Big Data Analytic, Text Mining and Machine Learning, Sentiment Analysis and Opinion Mining, Information Retrieval and Search Engine, Natural Language Processing, Software Engineering, Visualization)
- Supon Klaithin (Database Management System, Web Programming, Natural Language Processing)
- Santipong Thaiprayoon (Web and Database Management System, Information Retrieval and Search Engine, Natural Language Processing, Text Mining, Plagiarism Detection)
- Chatchawal Sangkeettrakarn (Information Retrieval, Search Engine, Question- Answering, Natural Language Processing, Mobile Application)
- Kanokorn Trakultaweekoon (Web Programming and Design, Database Management System, Visualization)
- Sittipong Saychum (Speech Synthesis Technology, Text analysis, Natural Language Processing, Embedded software development)
- Phuttapong Sertsi (Speech Recognition, Computer Programming, Mobile App Development)
- Nattapong Kurpukdee (Speech Synthesis, Text analysis, Computer Programming, Business Software Development, System and Database Administrator)
- Watthanasak Jeamwatthanachai, Ph.D (Smart Agent)
Contact
Speech and Text Understanding (STU)
Artificial Intelligence Research Group (AINRG)
National Electronics and Computer Technology Centre (NECTEC)
email: stu[at]nectec.or.th
Tel. : (+66) 02-564-6900 ext 2232