Upload
vishakhmarari
View
196
Download
4
Tags:
Embed Size (px)
Citation preview
Latent Fingerprint Matching Using Descriptor-Based
Hough Transform
1
Guided ByJose Martin M.J
Presented ByVishakh K.VRoll no: 61
INTRODUCTION
Law enforcement agencies are used since the early 20th century
Automated Fingerprint Identification System (AFIS)
A new AFIS is introduced for latent fingerprint matching which is notcurrently existing
2
3
TYPES OF FINGERPRINTS
Fig. 1. Three types of fingerprint impressions. Rolled and plain fingerprints are also called full fingerprints. (a) Rolled; (b) plain; (c) latent.
4
LATENT FINGERPRINTS
Lifted from surfaces of objects that are inadvertently touched or handled Usually smudgy and blurred, capture only a small finger area Large nonlinear distortion due to pressure variations
Fig. 2. Latent fingerprints of three different quality levels in NIST SD27.(a) Good; (b) bad; (c) ugly.
5
MINUTIA
Most important aspect in fingerprint analysisManually marked in latents Automatically extracted from rolled fingerprints
Fig.3. Fingerprint minutiae
6
LATENT MATCHING APPROACH
Fig. 4. Overview of the proposed approach
7
LATENT MATCHING APPROACH (Cont….)
A. Feature Extraction
1) Local Minutia Descriptor
Based on minutiae
Minutia Cylinder Code (MCC) – minutia based descriptor
Records neighbourhood minutia information as 3D function
Can be concatenated as a vector
Fig.5.(a) Latent and corresponding rolled print with a mated minutiae pair indicated(b) Sections of the cylinder corresponding to the minutia indicated in the latent and in the rolled print
8
2) Orientation Field Reconstruction
Minutiae based orientation field reconstruction algorithm is used
Estimates local ridge orientation in a block
Fig. 6. Latent fingerprint in NIST SD27 and the reconstructed orientation fieldoverlaid on the latent.
9
B. Alignment (registration)
Based on minutia matching Estimation of rotational and translational parameters Ratha et al. introduced an alignment which uses Generalized Hough
Transform Most similar minutia pair is used as base for transformation
parameters Our approach uses Descriptor-based Hough Transform (DBHT)
10
Parameter computation
Let be the minutiae sets,
To get efficient and accurate alignment,
1. voting using DBHT2. use of minutia pair that previously votes for a peak
11
ALGORITHM : Descriptor-based Hough Transform
12
ALGORITHM (Cont….)
13
C. Similarity Measure
For each alignment, a matching score between two fingerprints is computed
The minutiae matching score between the two fingerprints is given by
Where,denotes the similarity between the minutia cylinder codes of the ith pair of matched minutiae
maps the spatial distance of the ith pair of matched minutiae into asimilarity score
Take two values for Ts and mean of two matching score for two threshold aretaken
14
Fig. 7 (a)–(c) shows the latent, the true mate, and the rank-1 nonmate according to large threshold, respectively. (d)–(g) shows latent minutiae that were matched to rolled print minutiae in the following cases: (d) true mate using small threshold; (e) true mate using large threshold; (f) nonmate using small threshold; and (g) nonmate using large threshold. In (d)–(g), the scores corresponding to each case are
included.
15
Given the aligned latent orientation field and the rolled orientationfield , each containing k blocks, namely and , the similarity between the two orientation fields is given by
where, is 1 if both corresponding blocks are valid, and 0otherwise.
The overall matching score is given by
where the weight is empirically set as 0.4
16
Fig. 8. (a)–(c) show minutiae and the image of (a) a latent, (b) its true mate, and (c) the highest ranked nonmate according to minutiae matching. (d) and (f) show latent minutiae and orientation field (in blue)
aligned with minutiae and orientation field of the true mate. (e) and (g) show latent minutiae and orientation field (in blue) aligned with minutiae and orientation field of the nonmate.
17
EXPERIMENTAL RESULTS
Fig. 9. Performance of COTS2, MCC SDK, and Proposed Matcher when the union of manually marked minutiae (MMM) extracted from latents and automatically extracted minutiae by COTS2 from rolled prints is input to thematchers. (a) NIST SD27; (b) WVU LFD.
18
CONCLUSIONS AND FUTURE WORK
Presented a fingerprint matching algorithm using Descriptor Based-Hough Transform
Proposed system outperforms the well known commercial matchers
Scope of developing an indexing algorithm to speed upto include a texture-based descriptor to improve thematching accuracy
19
REFERENCES
A. A. Paulino, J. Feng, and A. K. Jain, “Latent fingerprint matching
using descriptor-based Hough transform,” in Proc. Int. Joint
Conf. Biometrics,
Oct. 2011, pp. 1–7.
Paulino,Feng,Jain Latent FP Matching Using Descriptor Based
Hough Transform_IJCB11
Wikipedia
20
21
THANK YOU