Optimal spoken dialog control in hands-free medical information systems
Jerzy Sas · Journal of Medical Informatics & Technologies · 2009
In the paper a method of optimal selection of utter ances used as command entry-words for voice controlled application is presented. Voice controlled programs seem to be par ticularly useful in the area of medical informatics , where a physician interacts with a program by voice while operating the medical device or being involved in examinations requiring manual activities. The proposed method selects command words from sets of proposals defined for each command so as to min imize the overall probability of incorrect command recognition. First the entry-word dissimilarity matrix is calculated. The word dissimilarities are evaluated using HMM models consisting of appropriately trained acoustic models of the phonemes constit uting words. The trained HMM is used as the sample utterance generator for t he word. The artificially created utterance samples are then recognized by speech recognizers created for pairs of words. The estimation of correct recognition probability is us ed as the word dissimilarity measure. The word dissimilarities are then used to determine the average assessment of words selection s that can be used as commands. Selection is created by choosing single word from sets of candidates defined for each comman d. Finally, suboptimal selection is found by using genetic algorithm. Expe riments carried out prove that suboptimal selection of command entry-words can observably increase the accuracy of spoken commands recognition in many cases.