Minutiae Verification and Classification for Fingerprint Matching

Raw image data offer rich source  of information for matching and classification.  For simplicity of pattern recognition system design, a sequential approach consisting of sensing,  feature extraction and matching   is   conventionally   adopted   where   each   stage   transforms   a   particular   component   of  information relatively independently. The interaction between these modules is limited.  Some of the  errors  in  the  end-to-end sequential  processing  can  be  easily eliminated especially for the feature extraction  stage by revisiting  the original  image data.  We propose a feedback  path for the feature  extraction stage, followed by a feature refinement stage for improving the matching performance. This performance improvement is illustrated  in the  context of  a minutiae-based  fingerprint  verification system. We show that a minutia verification stage  based on reexamining  the gray-scale profile  in a detected minutia's spatial neighborhood in the sensed image can improve the matching performance.