Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Stacked Denoising Autoencoder and Deep Neural Network Structure for Speech Lie Detection

An autoencoder, deep neural technology, applied in speech analysis, speech recognition, instruments, etc., can solve problems such as affecting the speech lie recognition rate

Active Publication Date: 2021-07-13
HENAN UNIVERSITY OF TECHNOLOGY
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Aiming at the problem that a single DNN structure affects the speech lie recognition rate, this paper proposes a structure combining stack denoising autoencoder and deep neural network

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Stacked Denoising Autoencoder and Deep Neural Network Structure for Speech Lie Detection
  • A Stacked Denoising Autoencoder and Deep Neural Network Structure for Speech Lie Detection
  • A Stacked Denoising Autoencoder and Deep Neural Network Structure for Speech Lie Detection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0016] To verify the performance of our proposed model, we conduct experiments on the CSC lie speech corpus. The CSC Lie Corpus is the first lie corpus designed and collected by language scientists. Study subjects were recruited into a "communication experiment" and told that the ability to succeed at deception represented certain desirable personal qualities, and the study attempted to identify who fit the 25 "top entrepreneurs" in the United States. These speeches are sampled at a rate of 16 kHz and divided into 5412 effective speech segments according to labels, including 2209 lies, and finally about 7h of lie speech samples are obtained. In this paper, 5411 speeches are cut out from the CSC library for experiments.

[0017] Step 1: After removing the parts with low sound quality, cut out 5411 voices from the library for experimentation. Each voice is about 2s long and contains 2209 lies voices. 4328 voices in the cut out voices are used as training set, and the remaining...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The feature that the existing voice lie detection algorithm often adopts is a combination feature, and the feature redundancy is relatively large. In view of this problem, the present invention discloses a stacked denoising self-encoder and deep neural network (SDAE-Deep Neural Network) for voice lie detection. DNN) structure. It consists of a two-layer encoding and decoding network followed by a DNN network. The structure first uses a two-layer denoising autoencoder structure to reduce feature redundancy. In order to prevent overfitting, dropout is added to each network layer of the stacked denoising autoencoder, and then a layer of DNN network is used for further learning. Features, and finally use the softmax classifier to fine-tune the network to obtain more representative features, thereby improving the network's ability to identify lies.

Description

technical field [0001] The invention belongs to the technical field of speech signal processing, and in particular relates to a stack-type denoising self-encoder and a deep neural network structure for speech lie detection. Background technique [0002] Psychologists have long been interested in human deception and its detection. Social psychology research has confirmed that lying is a common feature of daily social interactions, but people are not good at identifying lies. The identification of lies is of great significance for preventing telephone fraud, assisting criminal investigation cases and intelligence analysis, so the research on lie detection is a current research hotspot. [0003] In the field of speech lie detection, feature extraction and classification recognition are the core steps. At present, the features often used in speech recognition are combined features. Whether the features are effective depends largely on experience and luck, and the dimension of ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G10L19/012G10L15/02G10L25/30G10L25/51G06K9/62
CPCG10L19/012G10L15/02G10L25/30G10L25/51G06F18/213G06F18/2411G06F18/214
Inventor 方元博陶华伟傅洪亮雷沛之姜芃旭
Owner HENAN UNIVERSITY OF TECHNOLOGY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com