Detection of falsification using infrared imaging: Time and frequency domain analysis

摘要

Throughout the years many people have tried to master the art of lie detection; however no one was ever able to come up with a method that detects lies in more than 83% accuracy, which is non invasive, mobile, and cost effective. Therefore this article will attempt to show that by measuring temperature changes in the nose area, one could determine whether someone is lying or not better than current polygraph can. First, eleven different subjects were chosen to participate in this project. Then, using a FLIR ThermoVision A40 infrared camera the data was collected from eleven different subjects in three different sets of measurement. This paper presents the results using two different methods of detection. The first method of detection was using a time domain analysis and the second method of detection was using a frequency domain analysis. Each of these methods had an accuracy of 69% and 84% in detecting lies, respectively. Finally, the two methods were compared to current polygraph, which has an accuracy of 83%. It was concluded that it is possible to detect lies using an infrared camera and that it is actually better than current polygraph since it has higher accuracy, is non invasive, and cost effective.