A coarse-fine fingerprint scaling method

Fingerprint scaling is an important issue for fingerprints from different capture sensors. A good scale would process the resolution difference and somewhat the distortion of fingerprints. In this paper, we develop a coarse-fine method to estimate the optimal scale between the input and template fingerprints. The global scale is coarsely calculated based on the ridge distance map and finally determined by the histogram of local refined scale between all the matchable minutiae pairs. Experimental results demonstrate that our algorithm effectively improves the matching accuracy of the fingerprints from multi-type capture sensors.