Frontal infrared human face detection by distance from centroid method

Recently, thermal imaging has been receiving much attention owing to its application to face recognition in rapid mass fever detection. However, a non-frontal view of the face captured by the camera leads to erroneous temperature measurements. The paper proposes an automated thermal imaging system that is able to discriminate frontal from non-frontal face views with the assumption that at any one time, there is only one person in the field of view of the camera and no other heat-emitting objects are detected by the camera. The relationship between temperature and grey level values, various methods of enhancement, restoration, and segmentation of infrared images, and a way to determine the pose of a face are explained. A boundary signature, i.e. distance from centroid (DFC) of the human face to the lower part of the face outline is used. This measure has been used for the representation of objects, but it is now used to compare the degree of symmetry of the lower face outline. Experimental results are given to verify the validity of the method.