Analytical Survey of Computational Paralinguistic Systems for Automatic Recognition of Deception in Human Speech
Abstract
Introduction: Computational paralinguistics analyzes non-verbal aspects of human communication and speech, such as natural emotions, intonations, pronunciation features, speaker’s voice parameters, truth of a message, etc. The problem of automatic detection of truth/deception in spoken messages has importance in many practical applications. There are a great number of contemporary studies devoted to the development of software for automated systems of human speech deception detection. Purpose: We analyze and discuss the achievements and developments in the field of computational paralinguistics, particularly deception detection in human speech, in order to figure out the drawbacks of the available methods and define the ways to overcome them in developing a new automatic system. Results: The analysis of a wide spectrum of state-of-the-art scientific and technical literature discussing the results of the world-wide scientific research in this field for the last ten years, including International Computational Paralinguistic Challenge (ComParE) has shown that the researchers apply similar methods for deception/truth detection. However, the signal processing algorithms have some differences which can affect the accuracy of the deception recognition. We present a generalized scheme of a recognition system, its main components, as well as a classification of the most efficient methods used in the development of automatic systems for paralinguistic analysis of natural speech. At present, human speech deception detection has a lot of unresolved problems, both of technical and natural types, including taking into account individual features of a speaker (gender, age, emotional stability, national specificity, etc.). Overcoming these problems can significantly improve the system functionality.Published
2017-10-20
How to Cite
Velichko, A., Budkov, V., & Karpov, A. (2017). Analytical Survey of Computational Paralinguistic Systems for Automatic Recognition of Deception in Human Speech. Information and Control Systems, (5), 30-41. https://doi.org/10.15217/issn1684-8853.2017.5.30
Issue
Section
Information processing and control