fingerprint verification system
DESCRIPTION
Authentication. Fingeprint Image. Image Preprocessing. Fingerprint Image Enhancement. Minutiae Feature Extraction. Matching methods. Good quality Image. Good quality Fingerprint Image. Minutiae features. Database. Fingerprint Verification System. Fingerprint Segmentation. - PowerPoint PPT PresentationTRANSCRIPT
Fingerprint Verification System
Good quality Image
Good quality Fingerprint Image
AuthenticationFingeprint Image Fingerprint
Image Enhancement
Minutiae Feature
Extraction
Matching methods
Database
Minutiae features
ImagePreprocessing
Fingerprint Segmentation
Separation of fingerprint area (foreground) from the image background
• Traditional methods use block level features– Local histogram of ridge orientation– Gray-level variance– Magnitude of the gradient in each image block– Gabor feature
• My new method- point feature
Fingerprint Feature-Minutiae
Traditional Feature Detection Algorithm- Binarization-Thinning
– binarization followed by thinning step, the width of the ridges reduced to one pixel
– Location of minutiae points in the skeleton image • number of neighbor black pixels at a point of
interest in a 3 X 3 window• crossing number ( ending: cn(p) =1, bifurcation:
cn(p)=3, normal:cn(p) =2)– Thinning limitation: Aberrations and irregularity of the
binary ridge boundaries have an adverse effect on the skeletons, leads to the detection of spurious minutiae
New Minutiae Detection Method
Pout
Pin
Minutiae Point
Middle Point of SA and EB
(b) (c)
(a)
Pin
Pout
Pin × Pout
(d)
SA: Start Point of Pin
EB: End Point Pout
Pin × Pout
Figure 8 Minutiae Detection (a) Detection of turning points, (b) & (c) Vector cross product for determining the turning type, (d) Determining minutiae direction
Start
B
CF
Bifurcation
Post processing (Elimination of False Minutiae in the Image Boundary )
Determination of Turn Points• The ridge contours of fingerprint images can be consistently
traced in a counter-clockwise fashion
• Two types of turn points: left and right
• S(Pin, Pout) = x1y2 –x2y1
– Pin : Vector leading into the candidate point
– Pout: Vector leading out of the point of interest
– S(Pin, Pout) >0 indicates left turn, S(Pin, Pout) <0 indicates
right turn
– Significant turn can be determined by x1y1 + x2y2 < T
– Angle between Pin and Pout
IMAGE QUALITY MODELING -Proposed Limited Ring FFT Spectral Measures
the spectrum in polar coordinates, S(r, θ)
For each direction θ, Sθ( r ) – the spectrum behavior along a radial direction from the origin•For each frequency r, Sr(θ) – the spectrum behavior along a circle centered on the origin
Enhancement in High-curvature region of Fingerprint Image (2)
• Calculate the Gradients Gx, Gy• Calculate variances (Gxx, Gyy) and cross-
covariance (Gxy) of Gx and Gy• Calculate coherence mapsqrt((Gxx-Gyy)^2+4*Gxy^2)/(Gxx + Gyy)• Find the minimum coherence value in ROI• Add 0.1+ minimum (Coh)• Get the high curvature regions with region
property like centroid or bounding box
Enhancement Results
Enhancement resultsCore
Delta
Enhancement results