Post on 30-Dec-2015
Fingerprint Analysis(part 2)
Pavel Mrázek
Local ridge frequency
Local ridge frequency
Image enhancement / binarization• General rule:
– Smooth along ridges– Enhance ridge-valley contrast– Separate fingerprint from background
(segmentation)
• Various methods: – Convolution – PDEs – Morphology – Gabor filters – …
Gabor filters
• Several orientations• Several frequencies
• At each position,– select orientation– select frequency– filter using the appropriate Gabor filter
Gabor filters
Coherence enhancing shock filter
• Shock filter:
• Regularized:
Coherence enhancing shock filter
• Use direction estimate:w … dominant eigenvector of the structure tensor
Examples
Coherence enhancing shock filter
Ridge thinning• Thinning: morphological operation
• Pixel value set to background if ridge connectivity not affected
• Structuring element: typically 3x3 window
• 9 pixels, 512 possible configurations, look-up
Singular point detection• Methods for core and delta detection:
– Poincaré index– Irregularity of orientation field, curvature– Partitioning of orientation field
• Reliability problems
Texture features
Feature extraction summaryExtract features, store a template
• Prepare representation useful for matching– minutiae– …
• Reduce memory requirements(typical size 500 B – 30 kB)
• Privacy: fingerprint not stored
Enrollment• Register user, store data into a
database
Verification• Compare to enrolled template,
accept / reject a match
Identification• Recover identity, 1-to-N match
References• Maltoni et al.: Handbook of Fingerprint Recognition. Springer
2003.• Maltoni. A tutorial on fingerprint recognition. In LNCS 3161,
Springer 2005.• Hong, Wan, Jain. Fingerprint image enhancement: algorithm
and performance evaluation. IEEE PAMI 1998.• Zhou, Gu. A model-based method for the computation of
fingerprints’ orientation field. IEEE TIP 2004.• Weickert. Coherence enhancing shock filters. DAGM 2003.
• Contact: mrazekp -at- cmp.felk.cvut.cz