akanak
Newbie level 2
shivang patel
Hi,
I am currently studying on the Fingerprint Verification system (FVS) code (http://fvs.sourceforge.net/download.html) written in C to start up a toy project. I compiled the code in VC++ and applied some tricks to overcome some linux-to-VC modification problems. At last I obtained a code which is able to find minutia. The results of the feature extraction stage is not so bad but the matching and alignment part are still hard requirements for the code. I applied Madhav Kulkarni's implementation which was first proposed by Anil K. Jain et.al. paper (A. Jain, L. hong, R. Bolle, On-line fingerprint Verification, IEEE trans. on Pattern Analysis and Machine Intelligence, vol 19. No. 4, 1997) as a robust matching algorithm. But I guess this algorithm is useless unless you have two aligned fingerprints.
On the other hand the matching code which can be downloaded from the http://fvs.sourceforge.net/index.html (I upload the code within this message) may have some problems. are there any people who are faced with such problems? Because I could not get reasonable edit distances less than the specified threshold which is defined as # define EDIT_DIST_THRESHOLD (int) 10.
I also will be very pleased if you have any suggestions, codes, ... etc about fingerprint alignment that I can use it in this code. Maybe this will help other people to have a ridge-based FVS for the small projects..
Thank you for your interest.
sincerely
Hi,
I am currently studying on the Fingerprint Verification system (FVS) code (http://fvs.sourceforge.net/download.html) written in C to start up a toy project. I compiled the code in VC++ and applied some tricks to overcome some linux-to-VC modification problems. At last I obtained a code which is able to find minutia. The results of the feature extraction stage is not so bad but the matching and alignment part are still hard requirements for the code. I applied Madhav Kulkarni's implementation which was first proposed by Anil K. Jain et.al. paper (A. Jain, L. hong, R. Bolle, On-line fingerprint Verification, IEEE trans. on Pattern Analysis and Machine Intelligence, vol 19. No. 4, 1997) as a robust matching algorithm. But I guess this algorithm is useless unless you have two aligned fingerprints.
On the other hand the matching code which can be downloaded from the http://fvs.sourceforge.net/index.html (I upload the code within this message) may have some problems. are there any people who are faced with such problems? Because I could not get reasonable edit distances less than the specified threshold which is defined as # define EDIT_DIST_THRESHOLD (int) 10.
I also will be very pleased if you have any suggestions, codes, ... etc about fingerprint alignment that I can use it in this code. Maybe this will help other people to have a ridge-based FVS for the small projects..
Thank you for your interest.
sincerely
