A binaural sound source localization method and system based on deep neural network and convolutional neural network
A convolutional neural network and deep neural network technology, applied in the field of cascaded binaural sound source localization, can solve problems such as difficult to judge whether the sound source comes from the front half plane or the back half plane, confusion, and difficulty in extracting and characterizing the sound source. Achieve omnidirectional sound source localization, resist noise and reverberation, and avoid front-to-back confusion problems
Active Publication Date: 2022-02-18
PEKING UNIV SHENZHEN GRADUATE SCHOOL
View PDF6 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, it is difficult for traditional methods to extract clues that characterize the front and rear directions of the sound source, so it is difficult to judge whether the sound source comes from the front half plane or the back half plane, resulting in front and rear confusion.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0063] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. It can be understood that the described embodiments are only some embodiments of the present invention, not all of them. example. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts belong to the protection scope of the present invention.
[0064] This implementation example uses the binaural impulse response recorded with the Kemar artificial head, which has binaural impulse responses in different directions and distances within 360 degrees in the horizontal plane, and can simulate sound signals in different directions in the real environment.
[0065] The data collection process was carried out in a strict anechoic room, and the distance between all sound sources used and the c...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a binaural sound source localization method and system based on a deep neural network and a convolutional neural network. First, binaural energy differences and cross-correlation functions are extracted from binaural sound signals; then the extracted binaural energy differences and The cross-correlation function is input into the trained cascade neural network, and the direction of the sound source is judged in the deep neural network branch, and the detailed information of the sound source is extracted in the convolutional neural network, and the sound source is obtained in the front half plane of the robot head. Or the information of the second half plane; finally, the output results of the deep neural network and the convolutional neural network are integrated to obtain the final positioning result. The invention can resist the interference of noise and reverberation to a certain extent, completes the problem of judging the front, back and angle of the sound source, and realizes omnidirectional sound source positioning.
Description
technical field [0001] The invention belongs to the field of information technology, and relates to a binaural sound source localization method applied in speech perception and speech enhancement, in particular to a cascaded binaural sound source localization method and system based on a deep neural network and a convolutional neural network . Background technique [0002] The robot auditory system is one of the ways to make the interaction between robot and human, robot and the outside world more natural, convenient, efficient, friendly and intelligent. The robot auditory system mainly includes functions such as sound source positioning and tracking, speech enhancement, sound source separation and classification, speech recognition, speaker recognition, and speech emotion recognition. Among them, the robot's auditory sound source localization can provide auxiliary spatial information for other speech signal processing tasks. Therefore, whether it can accurately localize th...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G10L25/30G10L25/51G01S5/20
CPCG10L25/30G10L25/51G01S5/20
Inventor 刘宏庞程丁润伟吴璐璐孙永恒
Owner PEKING UNIV SHENZHEN GRADUATE SCHOOL