SPEECHLEX - phonological word modelling component of an experimental speech recognition system
نویسندگان
چکیده
SpeechLe1! is a module of an experimental speech recognition system. The main features of this system are the rule-based approach for the acousticphonetic detection, the incorporation of Fuzzy Logic to handle uncertain decisions, and an overall top-down recognition strategy. The task of SpeechLez is to set up a phonotactical network describing the words of a given vocabulary with common variations in pronunciation and to drive the acoustic-phonetic rules of another module. Processing is, in principle, carried out left-right, i.e. one syllable after the other, but inside-out within a syllable. Since the whole system is of an experimental nature, emphasis was laid on providing it with graphical aids to trace the processing steps and to support its development.
منابع مشابه
Phonetic Characterisation and Lexical Access in Non-segmental Speech Recognition
An isolated-word speech recognition system, built without the use of linear segments for acoustic modelling or lexical access, is justified, described and demonstrated. The system comprises phonetic feature analysis operating on four independent tiers, parallel phonotactic parsing, and lexical access based on a neural-network inspired lexicon structure. Performance is however still inferior to ...
متن کاملAllophone-based acoustic modeling for Persian phoneme recognition
Phoneme recognition is one of the fundamental phases of automatic speech recognition. Coarticulation which refers to the integration of sounds, is one of the important obstacles in phoneme recognition. In other words, each phone is influenced and changed by the characteristics of its neighbor phones, and coarticulation is responsible for most of these changes. The idea of modeling the effects o...
متن کاملFinite-state Transducer Base with Explicit Modeling of Ph
This article describes the design and the experimental evaluation of the first Hungarian large vocabulary continuous speech recognition (LVCSR) system. The architecture of the recognition system is based on the recently proposed weighted finite state transducer (WFST) paradigm. The task domain is the recognition of fluently read sentences selected from a major daily newspaper. Recognition perfo...
متن کاملFinite-state transducer based hungarian LVCSR with explicit modeling of phonological changes
This article describes the design and the experimental evaluation of the first Hungarian large vocabulary continuous speech recognition (LVCSR) system. The architecture of the recognition system is based on the recently proposed weighted finite state transducer (WFST) paradigm. The task domain is the recognition of fluently read sentences selected from a major daily newspaper. Recognition perfo...
متن کاملA model linking immediate serial recall, the Hebb repetition effect and the learning of phonological word forms.
We briefly review the considerable evidence for a common ordering mechanism underlying both immediate serial recall (ISR) tasks (e.g. digit span, non-word repetition) and the learning of phonological word forms. In addition, we discuss how recent work on the Hebb repetition effect is consistent with the idea that learning in this task is itself a laboratory analogue of the sequence-learning com...
متن کامل