Honor Terminal Co., Ltd. today announced a new patent for “Voiceprint Recognition Method and Electronic Equipment”, which can accurately recognize voices made by wearing a mask.
According to the patent specification, when the user wears a mask, the mask will inhibit the user’s voice to a certain extent, and the wake-up voice issued by the user wearing the mask is quite different from the registered voice, which may cause the electronic device to be unable to recognize the voice of the user. With the user’s own voice, the voiceprint recall rate is reduced, which affects the user experience.
In order to solve the above-mentioned technical problems, the present application provides a voiceprint recognition method and an electronic device, which improves the voiceprint recall rate and improves user experience.
The patent summary shows that the method includes:
performing feature extraction on the target speech to generate a first target spectrogram, and performing feature extraction on the registered speech to generate a second target spectrogram;
generating a target mask according to the first target spectrogram, and applying the target mask to the second target spectrogram to obtain a third target spectrogram;
extracting the first voiceprint vector from the first target spectrogram, and extracting the second voiceprint vector from the third target spectrogram;
If the similarity between the first and second voiceprint vectors is greater than or equal to the preset similarity threshold, it is determined that the target voice belongs to the user corresponding to the registered voice.
This patent application equates the masking effect of masks on speech to a mask acting on speech. This mask can flexibly and accurately simulate the suppression effect of masks on sounds of different frequencies so that it can more accurately identify the situation when the user is wearing a mask. The next voice, improve the voiceprint recall rate.
