Deception detection in text, speech and gestures using machine learning
Dissertant
Thesis advisor
Comitee Members
University
Al Akhawayn University
Faculty
School of Science and Engineering
Department
Software Engineering
University Country
Morocco
Degree
Master
Degree Date
2017
English Abstract
In legal systems worldwide, court trials decisions are made through a process that is carried according to applicable laws in order to prove defendants’ guilt or innocence.
Testimonies of the defendants and the witnesses are relevant to the decision-making process, and evaluating the honesty of the testimonies is critical for making accurate decisions.
Automating deception detection in court hearings could significantly affect trial decisions and requires a reliable system that can effectively and accurately identify deceptive behavior.
The identification of deceptive behavior is a topic that has been researched in several fields, such as psychology and criminology, which provided valuable information for understanding the differences between deceptive and truthful behavior.
This work explores the applicability of machine learning for detecting deception in real-life trial data.
The data used consists of real-life trial data recorded in the US for which the corresponding transcriptions and annotation of gestures is provided.
The evaluation of machine learning techniques in classifying hearing testimonies as deceptive or truthful is performed using text, speech and gestures, which correspond to the testimonies transcriptions, audio recordings, and annotated gestures respectively.
As part of this work, a comprehensive toolkit for detecting deception in audiovisual recordings was developed and includes the functionalities needed for building machine learning models from text, speech, and gestures and evaluating them.
The system developed provides a basis for evaluating deception detection using different datasets and configurations, as well as performing deception detection on audiovisual recordings, without technical knowledge, using existing models.
Main Subjects
Information Technology and Computer Science
Topics
No. of Pages
43
Table of Contents
Table of contents.
Abstract.
Abstract in Arabic.
Abstract in French.
Chapter One : Introduction.
Chapter Two : Background.
Chapter Three : Literature review.
Chapter Four : System features and architecture.
Chapter Five : System implementation.
Chapter Six : Evaluation and results.
Chapter Seven : Conclusion.
References.
American Psychological Association (APA)
Utarid, Husam. (2017). Deception detection in text, speech and gestures using machine learning. (Master's theses Theses and Dissertations Master). Al Akhawayn University, Morocco
https://search.emarefa.net/detail/BIM-775463
Modern Language Association (MLA)
Utarid, Husam. Deception detection in text, speech and gestures using machine learning. (Master's theses Theses and Dissertations Master). Al Akhawayn University. (2017).
https://search.emarefa.net/detail/BIM-775463
American Medical Association (AMA)
Utarid, Husam. (2017). Deception detection in text, speech and gestures using machine learning. (Master's theses Theses and Dissertations Master). Al Akhawayn University, Morocco
https://search.emarefa.net/detail/BIM-775463
Language
English
Data Type
Arab Theses
Record ID
BIM-775463
