Biometric identification in the moving maritime domain presents a number of unique challenges beyond the issues of long-range biometrics. Among these unique challenges is the issue of blur caused by both platform motion as well as the standard issues of atmospheric blur. It must also either provide for a controlled cameras system locking on and following faces, or address a larger operational issue, FOV that requires hardware-accelerated processing. We have previously shown that a moderately large sensor — at least 2 and ideally 3 or more MegaPixels — is needed to handle the variations in facial position caused by various sea states as well as demonstrating the feasibility of the core components of the FAMME system concept.
The core of the FAMME system is a combination of blur estimation combined with the output of multiple deblurred estimations, which is then fed into a recognition core. The ultimate goal is that the preprocessing, including motion estimation, deblurring, and face detection, will be implemented in FPGA or other embedded technologies. This project is a Phase II SBIR contract funded by the Office of Naval Research running from July 2008 to March 2010.