BRAIN-COMPUTER INTERFACE SYSTEM AND USER CONVERSATION INTENTION RECOGNITION METHOD USING SAME

摘要

The present invention relates to a brain-computer interface system and a user conversation intention recognition method using the same. The user conversation intention recognition method comprises: (a) a step of selecting an example word in a database storing sounds of N (N0) words, and outputting the sound of the selected example word to a user; (b) a step of measuring a brainwave or a soundwave created in an imagined speed, a speech action, or an audible action for the example word while the sound of the example word is outputted, and extracting and analyzing features between the measured brainwave data or soundwave data; (c) a step of learning a decoding model for detecting a word intended by a user by using the analyzed features; (d) a step of measuring a brainwave for a conversation sentence intended by a user in real time by a sensor module for detecting a brainwave of a user, and classifying candidate words estimated to a word intended by a user from the measured brainwave by priorities based on the learned decoding model; and (e) a step of combining the classified candidate words in accordance with priorities to generate a candidate sentence to provide a user with the candidate sentence in a voice or a text form. The decoding model includes: a classification model to classify a word intended by a user through features for the brainwave or the soundwave created in the imagined speed, the speech action, or the audible action; and a soundwave providing model to extract an envelope between the brainwave data and the soundwave data to derive a correlation, and then provide a sound waveform similar to a brainwave through the derived correlation.;COPYRIGHT KIPO 2020