C.-J. Lee and T.-N. Yang (Taiwan)
Minutiae detection, Gabor filters, principal Gabor basicfunctions, and fingerprint verification
On the features for fingerprint recognition, minutiae including ridge endings and bifurcations are the most reliable, especially for large-scale fingerprint databases. However, most traditional minutiae detection algorithms are complex and time-consuming. These processes always contain at least the following steps: image enhancement, orientation calculation, ridge detection, thinning, and minutiae detection. So many researchers developed some filter-based approaches to avoid above disadvantages, such as wavelet filters and Gabor filters. For small-scale fingerprint database, these filter-based approaches can match the requirements for recognition easily, but cannot for large-scale database. Therefore, this paper proposes principal Gabor basis functions to detect minutiae from raw fingerprint images directly. The proposed approach not only simplifies the processes of minutiae detection, but also preserves high recognition rates by using minutiae.
Important Links:
Go Back
