Enhancement of Biometrics with Advanced Pattern Recognition Techniques for Information Systems Applications

The Army is currently working to create effective biometric systems for positive identification of users under tactical conditions including poor capture conditions, quick response time requirements, narrow bandwidth. For instance, a soldier will need touse his weapon reliably even if the fingerprint image he supplies is rotated, smeared, occluded, or damaged. Many advances in pattern recognition have been made in other areas (computer vision, automatic target recognition) that can be applied to biometricmeasurement recognition. One such advance, the theory of quasi-invariants, studies features that are