ACOUSTIC EVENT DETECTION METHOD BASED ON DEEP LEARNING

Disclosed is a method that extracts a complex feature value included in the personal sound data based on the sound data including at least one sound source, classifies at least one or more sound source included in the complex feature value using an artificial intelligence model based on a fast regio...

Full description

Saved in:
Bibliographic Details
Main Authors KIM HONG KOOK, PARK IN YOUNG
Format Patent
LanguageEnglish
Korean
Published 02.06.2021
Subjects
ACOUSTICS
ALARM SYSTEMS
ELECTROPHONIC MUSICAL INSTRUMENTS
MUSICAL INSTRUMENTS
ORDER TELEGRAPHS
PHYSICS
SIGNALLING
SIGNALLING OR CALLING SYSTEMS
Online AccessGet full text

Cover

More Information
Summary:Disclosed is a method that extracts a complex feature value included in the personal sound data based on the sound data including at least one sound source, classifies at least one or more sound source included in the complex feature value using an artificial intelligence model based on a fast region-CNN-LSTM, and detects an event using the classified sound source. Therefore, the present invention is capable of providing improved real-time detection based on sound information as well as situation detection using image information. 본 개시는 적어도 하나 이상의 음원이 포함된 음향 데이터에 기초하여 사익 음향 데이터에 포함된 복합특징값을 추출하고, 고속 지역 합성곱 장단기 기억 신경망(Fast Region-CNN-LSTM)을 기반으로 하는 인공 지능 모델을 이용하여 상기 복합특징값에 포함된 적어도 하나 이상의 음원을 분류하고, 분류된 음원을 이용하여 사건을 감지하는 방법을 개시한다.
Bibliography:Application Number: KR20200035181