Facial Blood Vessels Activity in Drunk Persons using Thermal Infrared G. Koukiou*, V. Anastassopoulos † * † Electronics Laboratory, Physics Department University of Patras, Patras, Greece 26500 † vassilis@upatras.gr *gkoukiou@upatras.gr Keywords: Infrared Imaging, Face Recognition, Drunk Person Identification. Abstract In this work, the activity of the facial blood vessels of drunk persons is studied using thermal infrared images. Nonlinear anisotropic diffusion is applied to enhance the vessels on the images and after that top-hat transformation is used for isolating them from the rest information on the face. Simple thresholding is applied to raise the more active vessels. Registration procedures are employed to easily compare the vessel activity on the face of the drunk person with that on the sober person. In drunk persons, vessels around nose and eyes as well as on the forehead become more active. This work constitutes a preliminary study, which aims at qualitative and not quantitative results. Basically, the restricted number of the 20 persons that participated in the experiment is not considered adequate for statistical inference. 1 Introduction Biometrics is a very hot research area with a lot of publications the last few years. Applications are met in medicine, financial transactions, person identification and recognition and mainly in security issues. Research is carried out into several biometric problems [1, 2, 3], such as face and fingerprint recognition, facial expression classification and iris identification. Research in face recognition [2] has been biased towards the visible spectrum for a variety of reasons. Among those is the availability of low cost cameras in the visible portion of the electromagnetic spectrum and the fact that face recognition is one of the primary activities of the human visual system. However, machine recognition using visible light is difficult due to the fact that the acquired image changes with the conditions of illumination. Recently, emphasis has been given to acquiring information from faces in the thermal infrared spectrum [4, 5, 6, 7]. The main reason for this is that the temperature on various positions of the face depends on the physiological [8] as well as psychological [8, 9] conditions of the specific person, which in turn is strictly related to the distribution of the vessel network on it [4]. Some publications are available in the literature regarding drunk person identification [8, 11, 12, 13]. However, most of them are referred to automotive anti- drunk driving systems, which monitor the trajectory of the vehicle. Biosignals, which are electrical signals from the heart or brain, are used in [13]. In this paper, we present an approach to the problem of identifying a drunk person using facial vessels activity. In this approach we worked on thermal images. The approach was based on the fact that the facial vessels increase activity and some of them are brighter when the person consumes alcohol. For this purpose in each pair of IR images, which corresponds to a specific sober and drunk person, filtering and morphological operations were carried out to reduce noise and enhance edges. Specifically, anisotropic diffusion is applied to enhance the vessels on the images and after that top-hat transformation is used for isolating the vessels from the rest information on the face. Registration procedures are employed to reveal increased vessel activity on the face of the drunk person. Simple thresholding is applied to raise the more active vessels. In drunk persons vessels around nose and eyes become more active. In this study our intention is to show off, using Computer Vision methods, that a drunk person has different vessel activity compared to a sober person. No comparisons were carried out with persons being in bad psychological situation. It is actually a preliminary study, which examines the effect of a specific quantity of alcohol on persons, which are calm, and in normal psychological condition. On the other hand twenty students from the Electronics Laboratory were employed to consume alcohol. This sample is considered small, for statistical investigation, but it is quite suitable for a preliminary study. The total consumption of beer for each person was restricted to 4x330ml=1320ml. The organization of the paper is as follows. In section 2 a description of the data used is provided. The facial vessels are detected using filtering and top-hat transformation, which are described in section 3. In section 4, the activity of the vessels in the case of a drunk person is examined and