Pose Invariant Thermal Face Recognition Using AMI Moments
Naser Zaeri · 2016
Imaging in the visible spectrum demonstrates difficulties in recognizing the faces in conditions of varying illumination, especially under total darkness conditions. Further, the pose variations in such images add extra burden and heavy challenge on successful performance. As such, thermal face recognition has laid itself as a successful alternative solution and eventually has become an area of growing interest. In this paper, we present a new technique for thermal face recognition based on affine moment invariants (AMI) technique. AMI technique has become one of the most important shape descriptors. The technique will be implemented at the component level by dividing the face image into non-overlapped components. We anticipate that this approach will offer robustness against variability due to changes in localized regions of the faces. The new method will be tested on a new database comprising of images of different expressions with various severe poses, and were taken within different time-lapse. The experimental results have shown that the proposed technique offers high discriminability and performs efficiently, with Rank-1 successful rate of ~95% over the different poses.