SUMMARY OF SESSION 10 - Continous Speech Recognition II
نویسنده
چکیده
The session opened with two papers which addressed the compensation for noise and microphone characteristics. Significant performance improvement was demonstrated. Michael Picheny made an important comment in response to these papers, namely that these techniques have been developed before. Why are they not incorporated as a standard part o f speech preprocessing for all speech recognition systems?
منابع مشابه
Session 10: Evaluation of Systems on the Resource Management Database
This session presented papers on a number of different approaches to improving speech recognition performance, developed and evaluated on the DARPA resource management speech database. The session began with a summary presentation of results by Dave PalletL Most systems perform comparably, with word-level total error rates in the range of about 6 percent on the June 1990 RM2 speakerindependent ...
متن کاملSpoken Language Systems II
The papers in this session were concerned with higher-level processing in speech recognition systems and, in some cases, the interface between the speech-recognition and natural-language components of a spoken language system. The session consisted of talks from four DARPA sites, Dragon Systems, SRI International, BBN Systems and Technologies, and MIT Lincoln Laboratory. These talks were follow...
متن کاملSession 1: Spoken Language Systems
Without the ability to interpret natural language, speech recognition is suited only for a subset of tasks (though certainly not trivial ones), such as data entry, simple commands or dictation. Similarly, without speech recognition natural language is restricted to the interpretation of written language, a stylized form of human communication. Spoken language systems thus represent an attempt t...
متن کاملSpecial Session: Issues in a Language Processing (whe
The many aspects of audiovisual (AV) speech processing have attracted a loyal following from a wide range of perspectives and disciplines. Beginning with the landmark NATO workshop in 1995 [50], a series of international special sessions and meetings has fostered the development of numerous lines of AV speech research. This paper is an introduction to the special session and an inventory of sev...
متن کاملRecognition of Numbers by Using Demisyllables and Hidden Markov Models
A continous speech recognition system (called RAMSES) has been built based on the demisyllable as phonetic unit and tools from connected speech recognition. Speech is parameterized by band-pass lifted LPC-cepstra and demisyllables are represented by hidden Markov models (HMM). In this paper, the application of this system to recognize integer numbers from zero to one· thousand is described. The...
متن کامل