Hybrid recognition technology for isolated voice commands

Abstract

The paper deals with two elements of the artificial intelligence methods— the natural language processing and machine learning. Hybrid recognition technology for isolated Lithuanian voice commands is described. By the hybrid approach we assume the combination of two different recognition methods to achieve higher recognition accuracy. The method which is based on the machine learning algorithm to combine the recognition results provided by two different recognizers is described. The first recognizer was HTK-based Lithuanian recognizer, the second one—the Spanish language recognizer adapted to the Lithuanian language. The experimental results show that a hybrid decision-making rule learned by “random forest” classifier works with 99.46% accuracy and exceeds the accuracy of the “blind” decision-making rule (96.12%). The average hybrid operation accuracy reaches 99.24%, when the recognizer recognizes voice commands out of 12 known speakers, and is equal to 99.18%, when it is applied to the unknown speaker.