Speech technologies are being developed intensively in the recent years, especially the automatic\nspeech recognition as an additional input method in human interface and technical devices. Most of the\nknown algorithms for speech control have small probability of correct recognition. Widespread methods,\nlike Markov models and neural networks, which require large processing power, allow recognizing the\nwords with a probability of no more than 85ââ?¬â??92 %. Such accuracy is not enough to use the voice control on\nboard of a modern aircraft. The article is devoted to a problem of improving the automatic speech\nrecognitionââ?¬â?¢s accuracy. A version of word recognition algorithm based on the classical approach is\nsuggested, it includes the comparison with the patterns. In this work to improve the recognitionââ?¬â?¢s accuracy a\nnew method of calculating a similarity measurement between the recognizable word and the pattern, which\nbased on z-Fisher transformation, is described. This article also contains an algorithmââ?¬â?¢s modification that\ntakes into account the fixed ratios with the patterns of other words and uses the words adjustment to the\npattern with dynamic programming elements. The usage of fixed relations between words provides\nadditional information, which positively affects the recognition. The experimental results of the developed\nalgorithmââ?¬â?¢s approbation on a large amount of speech data are presented.
Loading....