Thermal Imaging of the Head for the Sensing and Identification of Concealed Intent

There are immediate needs for fast and cost-effective detection of concealed information or intent in homeline security applications such as border control, critical facility access. Unfortunately, traditional polygraph is not suited for this type of fast screening applications because it requires invasive contacts and extra time to administer. To meet today demand, new modalities have to be studied and examined for their potential efficacy in deception intent detection. Preliminary experiments showed that infrared image of face, or facial thermography, was a solution with promise. However, systematic studies are needed before it can be utilized in real applications. In this phase I project, we will address facial thermography problem in the following aspects: Guilty Knowledge Test (GKT) model design, data collection and verification, thermal feature selection and automatic recognition algorithm. GKT experiment is designed based on real world, personal experience and knowledge. The thermal data collected are compared and cross-validated against proven modality of brain activities as reflected in the P3 component of EEG in event-related potentials (ERP). The relevant facial thermagraphy features are selected through signal and statistical analysis. These features will be used in a dynamic pattern recognition scheme for fast and effective detection of deceptive intent. It is our objective that through our systematic and scientific research, the end results of this project will answer the following questions: which and how thermal facial features are related to brain activities resulted from deception related events; what is the range and limitation of these features; how effective they are in detecting deception.