Sparse spectrum signature extraction method for voice lie detection system

Abstract

The invention discloses a sparse spectrum signature extraction method for a voice lie detection system. The sparse spectrum signature extraction method for a voice lie detection system includes the following steps: 1) extracting a Mel frequency spectral coefficient and a wavelet packet frequency band cepstrum coefficient of a voice signal, and integrating the Mel frequency spectral coefficient with the wavelet packet frequency band cepstrum coefficient to form a cepstrum characteristic; 2) training the cepstrum characteristic by means of a K-singular value decomposition algorithm to obtain a mixed overcomplete representation dictionary; and 3) on the mixed overcomplete representation dictionary obtained in the step 2, utilizing an orthogonal matching pursuit algorithm to perform sparse coding on the cepstrum characteristic to obtain the sparse spectrum signature. The sparse spectrum signature extraction method for a voice lie detection system can make up the deficiency of providing medium and high frequency band information by a traditional Mel frequency spectral coefficient, and can also solve the redundancy problem of a non-linear fusion parameter set to reduce the calculating complexity of a classification model.

Description

technical field [0001] The invention belongs to the field of speech signal processing and pattern recognition, and in particular relates to a sparse spectrum feature extraction method used in a speech polygraph system. Background technique [0002] Lie detection technology has important applications in criminal investigation, military intelligence, and public place security inspections. It has broad development prospects and will gradually become a necessary technical detection method for the public security, judicial, personnel, and financial industries. When a person engages in lying behaviors such as deceiving others or intentionally creating false facts under the premise of strong motives (such as obtaining monetary benefits, evading guilt, etc.), the activity of the sympathetic nervous system or parasympathetic nervous system will be enhanced, and psychological Tension is created at the level of the liar, which in turn causes changes in the liar's heart rate, blood pres...

AI Technical Summary

Problems solved by technology

[0004] Feature redundancy is also an important factor affecting the accuracy of model recognition. As the types and dimensions of the extracted features increase, there will be redundant information in the data set. When performing model training and recognition of high-dimensional features, it will cause sample The problem of unbalanced quantity and sample dimensions and increased computational complexity makes it difficult to identify and distinguish the inter-class differences of data in the sample set

Traditional data dimensionality reduction methods such as Principal Component Analysis (PCA), Linear Discriminant Analysis (Linear Discriminant Analysis, LDA), etc., cannot handle nonlinear data sets well, and can only be used to deal with a Class features, cannot effectively remove those redundant and insignificant feature vectors

Images