VOICE RECOGNITION METHOD TO WHICH MULTI VOICE RECOGNITION MODULE IS APPLIED AND VOICE RECOGNITION DEVICE THEREFOR

摘要

The present invention relates to a voice recognition method to which a multi voice recognition module is applied and a voice recognition device therefor. Especially, the method applies keyword voice recognition and continuous voice recognition by stages to analyze a vocalized voice, thereby converting recorded data of a customer into texts to be used in a question file search, consultant management, and a marketing tool. In addition, multi voice recognition performance with regard to recorded data having low accuracy of voice recognition is improved so that the data can be used in analyzing a voice based on big data. Additionally, a language module using keyword voice recognition is selected and applied to continuous voice recognition to increase a voice recognition speed and performance for a variety of voice recognition services.;COPYRIGHT KIPO 2016