Problem Fingerprint recognision using Structure matching

[Umair.Mateen]

Hi i am working on Fingerprint recognition using structure matching,mean matching features associated with a center minutia and its neighbors.The problem is that its not working with rotated images. Actually after rotation many new minutia detects in the test image that causes problems. Even the old minutia features also change little bit(i.e. their distance from center,angle and other )
Kindly help me in this. Thanks

 

[mparvathi]

what algorithm do you use for pattern matching?

 

[Umair.Mateen]

I am mathcing features of Minutias. Actually i have made set of 5 minutia that are associated with one Center minutia.And for matching i first match the feature of center minuita then match feature of its neighboring 5 minutia (features of neighbors are ridge count,type,relative angle direction)

 

[tribu]

you could have it thru google..

 

[Umair.Mateen]

Actally problem is with preprocessor i am using
Segmentation
Normalization
Orientation estimation
Frequncy estimation
Gabor Filtering
Binirization
Thining

as pre-process steps.
I thing Gabor filtering causes problem on a rotated image. And when i apply thining on rotated image its each minutia point is differnt from straight image because of some new minutia poits are added