VOICE RECOGNITION APPARATUS AND METHOD OF RECOGNIZING THE VOICE

• Main Authors: LEE TAIHOON, SONG YOUJUNG
• Format: Patent
• Language: English; Korean
• Published: 27.03.2020
• Subjects: ACOUSTICS; CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

According to one embodiment of the present invention, a voice recognition device capable of increasing the accuracy for speaker identification comprises: an input unit receiving a voice; an extraction unit dividing the voice into phoneme units to extract a voice feature for each phoneme; and a control unit comparing the voice features with a reference voice feature and authenticating the voice based on the similarity. The extraction unit classifies a plurality of previously inputted reference voices into a plurality of groups reflecting an utterance state and extracts the reference voice feature from each of the plurality of groups. 본 발명의 일 실시 예에 의한 음성인식장치는, 음성을 입력 받는 입력부; 와 상기 음성을 음소 단위로 분할하여 각 음소별로 음성특성을 추출하는 추출부; 및 상기 음성특성을 참조음성특성과 비교하고 유사도에 기초하여 상기 음성을 인증하는 제어부를 포함하되, 상기 추출부는, 사전에 입력된 복수개의 참조음성을 발화 상태를 반영한 복수개의 그룹으로 분류하고 상기 복수개의 그룹 각각으로부터 상기 참조음성특성을 추출한다.