End-to-end recording equipment source identification method, identification system and computer equipment

Abstract

The invention belongs to the technical field of digital audio passive evidence obtaining, and discloses an end-to-end recording equipment source identification method, an identification system and computer equipment. The method comprises the following steps: extracting Mel-frequency cepstral coefficients in digital audio according to time sequence framing to serve as equipment source features, then dividing the equipment source features into time sequence feature segments, and extracting Gaussian mean matrixes respectively to obtain a time sequence Gaussian mean value matrix feature; performing deep representation learning on the time sequence Gaussian mean value matrix features by using a convolutional neural network, and extracting deep bottleneck features; and extracting time domain characteristics in the depth time sequence bottleneck characteristics through a bidirectional long-short-term memory neural network, and carrying out identification and classification on recording equipment sources. According to the method, a large number of recording equipment models can be effectively detected and distinguished, the particularity of digital audio is fully considered, and the accuracy and efficiency of an algorithm are improved; and when the source task of the recording equipment is carried out, the required voice data volume is small, each audio file only needs the length of several seconds, and non-voice segments do not need to be specially used for recognition.

Description

technical field [0001] The invention belongs to the technical field of digital audio passive forensics, and in particular relates to an end-to-end recording device source identification method based on temporal supervector representation learning. Background technique [0002] At present, in the past ten years, with the rapid development and progress of the electronic information industry, mobile devices such as smart phones and tablet computers have been widely popularized, and the convenience of audio recording, storage, and transmission has been greatly improved, and recording operations can be performed anytime, anywhere All can be done on mobile devices, no longer limited by the need to use professional recording equipment. [0003] When recording with a mobile device such as a smartphone, the recorded audio is saved as a digital audio data file. The digital audio data files obtained through recording carry a lot of information such as voice content, language type, spe...

AI Technical Summary

Problems solved by technology

[0013] (3) The existing non-end-to-end recording equipment source recognition model requires manual control of each stage in succession. There are many parameters that need to be manually adjusted, and the complexity is very high when the application is implemented, which makes it unable to be applied to other scenarios or other Identification of audio equipment sources, and the identification accuracy is low

Images