Adaptive source generator compensation and enhancement for speech recognition in noisy stressful environments

The author describes a low-vocabulary speech recognition algorithm which provides robust performance in noisy environments with particular emphasis on characteristics due to stress. A stressed speech source generator framework is formulated to achieve robust speech parameter characterization using a morphological constrained enhancement algorithm and stressed source compensation which is unique for each source generator across a stressed speaking class. An estimated source generator class sequence allows noise parameter enhancement and stress compensation schemes to adapt to changing speech generator types. A phonetic consistency rule is also employed based on input source generator partitioning. Average recognition rates for noisy stressful speech are shown to increase from an average 36.7% for a baseline recognizer to 74.7% for the new recognition algorithm. The new algorithm is also more consistent under varying noisy conditions as demonstrated by a decrease in standard deviation of recognition from 21.1 to 11.9, and a reduction in confusable word-pairs under noisy, stressed speaking conditions.<>