×
Apr 15, 2005 · The computational cost of a large-vocabulary continuous speech recognition system based on HMM is proportional to its vocabulary size.
Large-vocabulary continuous speech recognition using linear lexicon search and 1-best approximation tree-structured lexicon search. Authors: Norihide Kitaoka.
A search method using 1-best approximation treestructured lexicon search and linear lexicon search that achieves significant improvement of recognition ...
That is, we basically used the 1-best approximation tree-structured lexicon search and also used the linear lexicon search in parallel. ...
We propose a combinational use of linear lexicon and tree-structured lexicon in a 1-best approximation search algorithm for large vocabulary continuous ...
Page 1. Large-Vocabulary Continuous Speech Recognition Using Linear. Lexicon Search and 1-Best Approximation Tree-Structured. Lexicon Search. Norihide Kitaoka ...
– The pronunciation lexicon is structured as a tree. – Due to the ... • When using a (flat) linear lexicon and in the general n - gram case (n ≧2) ...
We propose a combinational use of linear lexicon and tree- structured lexicon in a 1-best approximation search algorithm for large vocabulary continuous ...
In this paper, we describe Large vocabulary Continuous Speech Recognition (LVCSR) system SPO-JUS++ which has been developed in our laboratory for over 20 ...
– The pronunciation lexicon is structured as a tree. – Due to the ... When using a (flat) linear lexicon and in the general n gram case (n ≧2) ...