Intoxicated Person Identification Using Thermal Infrared Images and Gait
Manas Kamal Bhuyan, Suchit Dhawle, Pradipta Sasmal, Γεωργία Κούκιου · 2018
In this paper, the activities of facial blood veins and temperature distribution and variations on the eye socket of drunk persons are studied using thermal images. Our proposed method considers walking pattern of a drunk person, which differentiate them from sober persons. For a sober person, vessels around the nose and eyes as well as nearer to the forehead region remain inactive and smooth, whereas these vessels become more active for a drunken person. Hence, thermal images can be used to detect drunkenness. The proposed method uses Curvelet Transform to capture the edges of a face to identify intoxicated candidate. The sclera and the iris are of the same temperature for the sober person. In contrary to this, the sclera temperature is high for intoxicated persons. SURF (Speeded up Robust Features) are used to accurately detect the temperature change in iris and sclera. The walking trajectory/pattern of a sober person and a drunk person can also be discriminated easily. Optical flow is employed to determine motion trajectories of drunk and sober persons. Finally, classification is done by using Random Forest and the Support Vector Machine. The experimental results show the efficacy of the proposed method.