Thermal imaging for face recognition in optical security systems

摘要

Face recognition belongs to the most fundamental human perceptual capabilities. Today, the technology is encapsulated in commercial software that runs on any low cost personal computer and is capable to detect the human faces anywhere in a scene and rapidly identifying them by matching the face against its database of known faces. This paper discusses the possibility of exploiting of the registration of thermal images and theirs computer aided processing for facial recognition systems. The infrared camera captures the pattern of heat distribution emitted by the body sufaces. The captured thermogram is read, digitized and stored by a computer for further analysis. Then the computer performs two distinct but related tasks: locating the face within the image and its identification. A new image taken by an infrared camera can be compared to a previously stored thermogram in order to recognize or verify the identity of an individual. Before any recognition can take place, the face must first be located within the digital image. Identification can consist of complex minirecognition algorithms that find distances between basis points. Thermal recognition of human faces is light independent and non-intrusive. The identity can be verified without contact, in low light or even in the absence of light. thermal recognition can be integrated into the existing security systems for the following uses: access control and computer security.