Fingerprint matching incorporating ridge features with Adaptive Coarse filter and Thin Plate Spline Warping (TPSW) model

The performance of a fingerprint matching system is affected by the distortion introduced in the fingerprint impression during feature extraction process. This fingerprint features, such as ridge curves distorted in a complex manner. With the recent advancements in fingerprint sensing technology, many sensor devices to capture the fingerprint are now equipped with probability of dual resolution scanning capability. Anyway, only increasing the scan resolution does not provide any improvement towards performance in fingerprint matching, unless the distortion is minimized and an optimized way of improving the matching process is evolved. In our research work, we plan to present a technique to minimize the distortion level in fingerprint based on ridge curve and an optimal model based on the consistent distortion which results in improved performance. The matching process consists of three main modules. The first module extracts the indispensable information from the reference image, the second module provides with the information obtained from the reference image by Adaptive Coarse filtering algorithm. Finally the third module performs and validates the actual matching process which results in optimality and minimizes the distortion level using Thin Plate Spline Warping algorithm (TPSW). Ridge feature are extracted using Adaptive Coarse filters and are matched using the Thin Plate Spline Warping (TPSW) algorithm. The proposed matching model based on ridge features lead to a better alignment of fingerprint images compared to matching model based on pores and ridges. Results reported on the NIST Special Database 14 being distributed for use in development and testing of automated fingerprint feature extraction and matching process on a set of images demonstrate the effectiveness of matching process. There is a relative reduction of 30 percent in the equal error rate (EER) of the matching system when adaptive coarse filter and TPSW are employed. Results reported on the NIST Special Database 14 reference set, featuring 90 percent correct verification (10 percent false negative) and 0.05 percent false positive, demonstrate the effectiveness of the matching process. On the other hand, there results in a relative reduction of 30 percent in the equal error rate of the matching system. This significant improvement results in an improvement in the matching performance.