Evaluation of minutia cylinder-code on fingerprint cross-matching and its improvement with scale

摘要

Recognition of fingerprints from multi-type capture sensors is a challenging work. While minutia cylinder-code, which is suggested to be invariant for translation and rotation and robust against skin distortion and small feature extraction errors, is an excellent feature proposed recently. In this paper, we conduct a series of experiments with minutia cylinder-code to evaluate its performance on fingerprints from multi-type capture sensors. Then a scale step is added to deal with the different resolutions and the creation of minutia cylinder-code is modified accordingly. The modified minutia cylinder-code with scale and two other matchers are tested. The experimental results proved that minutia cylinder-code is an effective and efficient feature for fingerprints from multi-type sensors, especially after modified with scale.