The Improved Fingerprint Ridge Distance Extraction Algorithm Based on Embedded Environment

摘要

For the problem that the traditional ridge distance estimation method is influenced by the quality of fingerprint images, and high computational complexity, so it is not suitable for embedded environment, the paper proposes a new method of ridge distance estimation. The algorithm firstly select candidate image block according to the direction of curvature, Secondly we execute the second candidate selection by using variance and mean based on the above. Finally we calculate the ridge distance in representative image block by using a statistical window method. Experiment is carried out in the FVC2004 database, and it has shown that the representative image blocks can also be used to reliably estimate the average ridge distance.