Detection of Intoxicated Person using Thermal Infrared Images

The proposed method helps in the identification of an intoxicated person from thermal images. A significant change in the body temperature and irregular blood flow after consumption of alcohol generate some textural changes along the areas of eyes, face, hands, and ears of a person. Textural changes help in identifying temperature distribution along the different regions of the face and other parts of the body. Different changes in facial blood veins can be observed as well. For this paper, Non Sub-sampled Contourlet Transform (NSCT) is used to extract the texture features from thermal images of eye, face, hand, and ear. An unsupervised feature selection technique is utilized for feature optimization which is followed by classification using Support Vector Machine. Extensive experiments prove the efficiency of the proposed work.