Add your own alternative version
Stats
1.2M views 124.4K downloads 658 bookmarked
Posted
30 Jul 2010
|
Comments and Discussions
|
|
I was tried to use algorithm JY on your framework but got error when run at below.
Please help.
throw new InvalidOperationException("Unable to extract JYFeatures: Unassigned skeleton image extractor!");
var fingerprintImg1 = ImageLoader.LoadImage(fileName1);
var fingerprintImg2 = ImageLoader.LoadImage(fileName2);
var featExtractor = new JYFeatureExtractor() { MinutiaListExtractor = new Ratha1995MinutiaeExtractorFingerprint() };
var features1 = featExtractor.ExtractFeatures(fingerprintImg1);
var features2 = featExtractor.ExtractFeatures(fingerprintImg2);
var matcher = new JY();
double similarity = matcher.Match(features1, features2);
score = similarity.ToString("0.000");
Console.WriteLine("the matched score is {0}", score);
|
|
|
|
|
Hi:
I'm sorry for taking so long to reply. This is what is missing with your code:
var featExtractor = new JYFeatureExtractor()
{
MinutiaListExtractor = new Ratha1995MinutiaeExtractorFingerprint()
SkeletonImgExtractor = new Ratha1995SkeImgExtractor()
};
I suggest you download the current version of our framework from https://sites.google.com/site/miguelmedinaperez/software/fprframework[^]
Please, also notice that the performance of any matcher highly depends on the quality of the minutiae. Hence, we suggest you extract the minutiae using a better algorithm and then use our framework based on those minutiae.
Best regards,
Miguel Angel Medina Pérez, PhD
|
|
|
|
|
Hey, very nice job on the code. But i was wondering what kind of fingerprint scanners are you guys using? I am currently doing a final year project but have no idea what to get, something efficient and not too pricy. Would love your recommendation. Thanks in advance 
|
|
|
|
|
Hi:
We are not using any scanner. Our framework allows matching fingerprints but it includes no driver to connect to a scanner.
You should get some scanner that comes with some SDK so you can create your own code to connect to the scanner.
P.D. I suggest you download the last version of our framework from: https://sites.google.com/site/miguelmedinaperez/software/fprframework[^]
Best regards,
Miguel Angel Medina Pérez, PhD.
|
|
|
|
|
Sir(s),
First of all let me express how thankful I am to you all who contributed in this project. I actually have a couple of requests. Firstly, Will you please, allow me to use your framework in my project(M.Tech)? and Secondly, I am mere a newbie in the arena of programming in c#, and even more neonatal in the context of Fingerprint matching, so it would be very helpful for all of us(at least like me), who want to rework/understand the working mechanism/principle of your framework, to have a lecture/'video tutorial', explaining the show behind the scene.
You must be very busy, but sir, please consider my request and consider how good will it be if a good article like this helps programmers to better understand 'fingerprint matching' and make it a cake-walk, turning this a standard(universally accepted) .NET framework for fingerprint matching at a much larger scale, via a tutorial(Documentation).
Thank you so much for your time.
P.S. : Please share the link of video tutorial/Documentation on this project(If you have made/make any) .
|
|
|
|
|
Hi:
The section "Matching fingerprints out of this framework" contains the necessary information to use our algorithms in your project.
Best regards,
Miguel Angel Medina Pérez, PhD
|
|
|
|
|
Thank you So much, I'll give it a try and contact you in case of any help required. 
|
|
|
|
|
Hi !
At the moment there not enough meat to make a proper article but i made a huge refactoring of your project, fixing some problems with the reflection and improving performances.
I really want your opinion
GitHub - matteofabbri/fingerprints[^]
Thanks, bye
|
|
|
|
|
|
Sorry for very late response, i didn't saw you message.
Sadly no, i built it from the code attached to the article but i will be very happy to upgrade to the latest version 
|
|
|
|
|
That will be great. I suggest you creating a webpage, or publishing a document, listing the things that you refactored and explaining how the research community could now use this new version of the fingerprint recognition framework.
Best regards,
Miguel Angel Medina Pérez, PhD
|
|
|
|
|
I have been trying to Debug and Build the programme, but it keeps giving an error message that this file "FR.FMExperimenter.exe is missing. And when i searched through the extracted zip folder, it was not there. please need your help.
|
|
|
|
|
Hi:
I suggest you use the most recent version of our framework (https://sites.google.com/site/miguelmedinaperez/software/fprframework).
Best regards,
Miguel Angel Medina Pérez, PhD
|
|
|
|
|
I want to ask about the way of performing the experiments using a different database like (livDet).
how can I add such database?
Thank you for your helping.
|
|
|
|
|
Hi:
You must create your evaluation protocol because each database contains its format regarding file names, amount of genuine matching fingerprints, amount of imposter matching fingerprints, etc.
You can copy one of the evaluation protocols that we included in the framework, and then adapt it so it can be used with your database.
Best regards,
Miguel Angel Medina-Pérez, PhD
|
|
|
|
|
Thank you for your valuable reply.
I'll try to do it.
Best Regards
|
|
|
|
|
Hello
first of all I am grateful for all efforts that you present, but if you can make video about how integrate all part of this framework because I am very confused and can not use it until now, when I started run it it dosen't work .
thank you so much
|
|
|
|
|
Hi:
I'm glad you liked our work. I suggest you use the most recent version of our framework (https://sites.google.com/site/miguelmedinaperez/software/fprframework).
Best regards,
Miguel Angel Medina Pérez, PhD
|
|
|
|
|
Hi,
Any sample code for web application integration fingerprint ?
|
|
|
|
|
Hi:
We don't have any web integration solution.
Best regards,
Miguel Angel Medina Pérez, Ph.D.
|
|
|
|
|
Thanks a lot for making it open source!
It was the most elaborate framework that I had come across. The documentation helped me to run the references on my custom application successfully. And I am able to get the level of similarity between different prints (I used the standard Mtriplets and Ratha1995 for testing).
I kindly need your suggestions on going ahead to use the same for Biometric Attendance system.
What should I keep as the cut off similarity for deciding that the match is accurate (Please note that I found similarities as high as 0.2 between thumb impressions from different persons.)
Also how can I use multiple images for each person in database and use the same for comparing at runtime.
Will average of the similarities with the multiple database images give a meaningful result?
Best Regards,
Vignesh Sai
|
|
|
|
|
Dear Vignesh:
I'm glad that our framework was useful for your application. Here it goes the answer to your question:
1) And I am able to get the level of similarity between different prints (I used the standard Mtriplets and Ratha1995 for testing).
Answer: We implemented Ratha1995 for education purpose. You can obtain better features by using another free framework (SourceAFIS sdk). We created a new version of our framework (FPRFramework - Miguel Angel Medina Pérez[^]) which includes a wrapper of SourceAFIS SDK in order to use it withing our framework.
2) What should I keep as the cut off similarity for deciding that the match is accurate (Please note that I found similarities as high as 0.2 between thumb impressions from different persons.)
Answer: You should create your own database using the sensor that you intend to work with. Then, you should perform an evaluation of the matching algorithm on your custom database using a protocol such as the one used in FVC2000, FVC2002, FVC2004, and FVC2006. Once you evaluate the algorithm, you may select a threshold (a cut off similarity value) from the ROC curve. Our framework includes those evaluation protocols for FVC2002, FVC2004 and so on; so you will only need to adapt those codes to the size of your database. In order to understand ROC curves in order to determine your threshold, you will need some background knowledge on fingerprint recognition; so I recommend you reading the "Handbook of fingerprint recognition."
3) Also how can I use multiple images for each person in database and use the same for comparing at runtime. Will average of the similarities with the multiple database images give a meaningful result?
Answer: Yes, the average will work. You could read the "Handbook of fingerprint recognition" for more elaborated strategies.
Best regards,
Miguel Angel Medina Pérez, Ph.D.
|
|
|
|
|
hola octavio y demas autores, en mi empresa estan buscando solo reconocer que en los documentos jpg que estan almacenados en una base de datos, reconocer las imagenes de huella y recortar esa parte de la imagen, con el framework que lineas de codigo serian para reconocer que si existe una huella en un documento? lo demas seria recortarla... 
|
|
|
|
|
Hi Greg:
Thanks for your support. Our framework does not include any specific code to do what you want. Such a thing would require more work. Write to my personal email (miguelmedinapere@gmail.com) for further assistance.
Best wishes,
Miguel Angel Medina-Pérez, Ph.D.
|
|
|
|
|
A great work, the most interesting article I ever read , thank you.
I'm new to this field and I need help here, how can I classify a given fingerprint into one of the main four or five, I don't know the exact number, defined classes (left-loop, right-loop, whorl...).
Is there a method in your framework to achieve that. If no, do you have a reference that help me to do that.?
Please help.
Thanks
|
|
|
|
|
|
General News Suggestion Question Bug Answer Joke Praise Rant Admin
Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.
|
- Best C# article of July 2010
|