A fingerprint verification algorithm using tessellated invariant moment features

Effective enhancement of low-quality fingerprints with local ridge compensation

“…To illustrate the performance of the proposed preprocessing algorithm, a fingerprint verification system, which uses tessellated invariant moment features [7], was tested with the preprocessing algorithms. We used the Gabor-filter enhancement [2], the STFT enhancement [4], and our proposed enhancement method for comparison.…”

“…Fingerprint recognition algorithms are usually classified into two classes: minutiae-based and image-based [7]. The former primarily uses information on minutiae for recognition and the latter uses a reference point along with a number of attributes of ridges.…”

A New Approach to Fingerprint Detection Using a Combination of Minutiae Points and Invariant Moments Parameters

“…The overall performance of AFIS mostly relies on extracting features and its efficient matching mechanisms. It still deserves a much more advanced and accessible high volume fingerprint database that takes into account the computational speed and accuracy of matching [5,6]. To extract and compare of an image for matching, roughly three techniques can be categorized: minutiae-based, correlation-based, and hybrid explained in [7].…”

“…All of the above methods have more or less several imperfections. Some of these types of imperfections include: the performance of an algorithm and its computational speed degrades when the image quality is not satisfactory, for example the case of a multi-spectral noisy image [5]; a limited efficiency when a different scale position and rotation angle are used for the same input images [15]; or suffering from time-consuming alignment (correlation-based).…”

“…K. Jain et al [8] proposed a correlation-based method where the changes in position of the input image are skipped over with respect to a reference point. In [5], J. C. Yang et al extracted invariant moment properties on a certain Region of Interest (ROI) and an eigen value weighted cosine distance was used to detect similarities or dissimilarities. J. Garcia et al [16] applied a filtering method and a moment feature to calculate the parameter for matching.…”