Speech enhancement method and device based on dual-channel neural network time-frequency masking, and hearing-aid equipment

Abstract

The invention relates to the technical field of hearing-aid speech enhancement, and particularly relates to a speech enhancement method and device based on dual-channel neural network time-frequency masking and hearing-aid equipment. The method comprises the steps: receiving a speech signal through employing two microphones under noise and reverberation conditions; performing preliminary speech enhancement on each path of received microphone signals through the trained single-channel neural network, and removing the noise in the same direction as the target speech; different from traditional positioning which blindly depends on signal energy, enabling the self-adaptive ratio mask to provide a view of a target signal, accurately identifying a speech dominant time frequency unit on each microphone channel, using the time frequency units for sound source positioning, and ensuring that high positioning precision is obtained under noise and reverberation conditions; and inputting the calculated weight into a WPD beam former to remove noisy speech in different directions from the target speech and suppress room reverberation so as to obtain enhanced speech with good speech quality and high intelligibility.

Description

technical field [0001] The invention relates to the technical field of hearing aid speech enhancement, in particular to a speech enhancement method, device and hearing aid equipment based on dual-channel neural network time-frequency masking. Background technique [0002] Hearing-impaired patients have reduced speech recognition ability due to damage to their auditory organs, making their speech comprehension in a noisy environment much lower than that of normal-hearing people. Hearing loss can cause great inconvenience in life. Temporary and permanent hearing loss is becoming younger as people become more exposed to portable music players, according to the National Health Survey. Wearing hearing aids is one of the safest and most effective ways to solve the problem of hearing loss, and it has an irreplaceable role in improving the hearing of people with hearing loss. Speech comprehension of patients with hearing loss in noisy environments is one of the key factors affecti...

AI Technical Summary

Problems solved by technology

[0007] In view of the technical defects and technical drawbacks in the prior art, embodiments of the present invention provide a speech enhancement method, device, and hearing aid device based on dual-channel neural network time-frequency masking that overcome the above problems or at least partially solve the above problems. It solves the problems that the performance of the existing hearing aid speech enhancement algorithm drops sharply in a noisy environment, the ability to suppress directional noise is poor, and the room reverberation is not considered at the same time, resulting in low intelligibility of the enhanced speech

Images