In this paper continuation of previous research done on lie detection using infrared imaging has been presented. Previously it was shown that lie detection using infrared imaging is possible and is actually very accurate (with accuracy of up to 84%). This paper showed a different type of detection algorithm using the implementation of time moving average. All data for this experiment was collected from 11 subjects, which were asked to answer 15 questions in three different times. The first time they were asked to state the truth, the second time they were asked to lie, and finally they were asked to mix their answers and state when they lied and when they said the truth. Then three different post processingmethods were used in order to attempt and detect lies.
This paper presents the results of detection using the time moving average method. It was first shown that this method has the same characteristics that support the hypothesis, meaning the truth data does not have wide peaks, and the lie data has wide peaks. Then the accuracy of the time domain moving average was found to be 85%, which is greater than current polygraph and previous methods presented. Finally, it was concluded that this method is more efficient and these findings could lead to a new era in the lie detection industry due to this method’s efficiency, accuracy, and cost effectiveness.