Multi-Spectral Face Recognition - Fusion of Visual Imagery with Physiological Information

We present a novel multi-spectral approach for face recognition using visual imagery as well as the physiological information extracted from thermal facial imagery. The main point of this line of research is that physiological information available only in thermal infrared, can improve the performance and enhance the capabilities of standard visual face recognition methods. For each subject in the database, we store facial images collected simultaneously in the visual and thermal bands. For each of the thermal images, we first delineate the human face from the background using the Bayesian framework. We then extract the blood vessels present on the segmented facial tissue using image morphology. The extracted vascular network produces contour shapes that are unique to each individual. The branching points of the skeletonized vascular network, referred to as thermal minutia points (TMPs), are an effective feature abstraction. During the classification stage, we match the local and global structures of TMPs extracted from the test image with those of the corresponding images in the database. We fuse the recognition results of our thermal imaging algorithm with those of a popular visual imaging algorithm. We have conducted experiments on a large database of co-registered visual and thermal facial images. The good experimental results show that the proposed fusion approach has merit and promise.