The development of a wireless neuro sensor-based lie detector

Norman, Nurnajwa Amani (2021) The development of a wireless neuro sensor-based lie detector. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) The development of a wireless neuro sensorbased lie detector.pdf - Submitted Version Download (331kB)
	Text (Full Text) The development of a wireless neuro sensorbased lie detector.pdf - Submitted Version Restricted to Registered users only Download (2MB)

Abstract

Technology nowadays are growing rapidly. This helps to increases the motivation for the community to be able to produce and continue new innovations. The new innovations and technology produces is known as electroencephalogram (EEG) where this innovation is used to detect brain signals. This thesis uses the non-invasive measurement method where the electrical signals from the brain will be obtained by the placement of multiple electrodes on the forehead. Lie is an act of covering up something that only those who lies knows the correct situation or statements. The advancement of cognitive science and neuroscience EEG analysis gives a better understanding of brain function. An Artificial Neural Network (ANN) is used as a machine learning technique to analyses the EEG signals. There are two types of data uses in this thesis which is dataset from Neurosky MindWave while real data that are gained form Neurosky MindLink Sensor. In real data, the Guilty Knowledge Test (GKT) method is used as a question to ask to the subject where it consists of three types of questions which are the truth questions, the lie questions and finally the baseline questions. Dataset is gained from Neurosky MindWave sensor by inviting subject in playing game card so that the process is done naturally. The sensor used for this project is Neurosky MindLink EEG Sensor where MATLAB software is used to process the data achieved. Next, Standard Deviation Method is being chosen as an input for training neural network of this thesis while Artificial Neural Network (ANN) is used as a classifier to analyses the EEG signals. The results achieved is when the subjects are lying, it will show and index of 2 while when a subject is telling the truth, it will show and index of 1. The accuracy of dataset is 95% based on results of confusion matrix in classifier while for real data, the accuracy achieved is 80% based on the results obtained.

Actions (login required)

View Item