Vocal print identity sameness identifying method and device, and computer readable storage medium

Abstract

The invention provides a vocal print identity sameness identifying method and a vocal print identity sameness identifying device, and a computer readable storage medium. The method comprises the following steps: acquiring to-be-identified audios used for sameness verification; carrying out resonance peak extraction on the to-be-identified audios, thus obtaining a broadband vocal print chart containing the resonance peak trend of the to-be-identified audios, wherein the horizontal coordinate of the broadband vocal print chart is the bandwidth, and the vertical coordinate is the frequency; in the broadband vocal print chart, determining the characteristic parameters among the corresponding resonance peaks of the to-be-identified audios; and judging that whether the characteristic parametersare within the preset range or not, wherein if so, the result passes the sameness verification, otherwise, the result does not pass the sameness verification. According to the technical scheme, by displaying the vocal print information of the audios into the broadband vocal print chart with the horizontal coordinate being the bandwidth and the vertical coordinate as the frequency, the vocal printcharacteristic of the audios can be intuitively displayed, then the vocal print characteristic and identity sameness can be rapidly verified, and therefore, the vocal print identifying efficiency is improved.

Description

technical field [0001] The present invention relates to the technical field of speech recognition, in particular to a voiceprint identity identification method, device and computer-readable storage medium. Background technique [0002] Voiceprint identification, also known as voice identification, is a specialized technology for personal identification of spoken words. Through voiceprint identity identification, it is possible to accurately determine whether several pieces of audio are from the same person. Therefore, it can be applied to many scenarios , such as the identification of criminal suspects in justice. [0003] In the existing voiceprint identification, the abscissa of the conventional two-dimensional spectrogram (voiceprint broadband map) is time, the ordinate is frequency, and the color depth of the coordinate points indicates the sound intensity, which is difficult to quantify. Voiceprint features such as peak trends are not intuitive enough, and a lot of lab...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a voiceprint identification method, device, and computer-readable storage medium, which are used to solve the cost of traditional voiceprint identification due to the invisibility of the voiceprint features of the two-dimensional spectrogram. A lot of labor costs to observe the technical problems of the comparison

Images