Detecting the "ngerprint minutiae by adaptive tracing the gray-level ridge

This paper presents a minutiae detection procedure based on adaptive tracing the gray-level ridge of the "ngerprint image with piecewise linear lines of di!erent length. The original "ngerprint image is smoothed with an adaptive-oriented smoothing "lter only at some selected points. This will greatly reduce the computational time. Each ridge in the skeleton is labeled with a number so that each detected minutia is associated with one or two ridge numbers, which is useful for post processing. We objectively assess the performance of this approach by using two large "ngerprint databases. 2001 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.