Fingerprinting

helper · Post by **helper** » Sun Aug 20, 2006 3:08 am

I have changed my topic to "How hard is it to fake a fingerprint?"

The experiment will be from this website but I will put in more kinds of moulds.
[url]http://cryptome.org/gummy.htm[/url]

So if the fingerprint scanner accpets most of the artificial fingers, that means it will be easy to fake a fingerprint.

Does anyone think this experiment is good?

deleted-2131 · Post by **deleted-2131** » Mon Aug 21, 2006 5:07 pm

The topic of your science project is just fine; however, your specific question "How hard is it to fake a fingerprint?" is not testable. A good science fair question is one that is testable. I would reccomend rewording your question to make it testable. For example, you might use your artifical "fingers" on different fingerprint scanners and see how often the scanner is fooled. A questions could be something like "How often can fingerprint scanners be fooled by artificial fingerprints?"

If you have any more questions, please let me know. I'd be more than happy to help you design and carry out your project.

helper · Post by **helper** » Mon Aug 21, 2006 9:11 pm

"How often can fingerprint scanners be fooled by artificial fingerprints?" is exactly what I mean and the same experiment. But isn't that practically the same thing as "How hard is it to fake a fingerprint?"

deleted-2131 · Post by **deleted-2131** » Tue Aug 22, 2006 6:27 pm

Yes, your re-worded question is just as good (if not better) than the suggestion I made. I would reccomend using your new wording.

Great job!

helper · Post by **helper** » Wed Aug 23, 2006 3:37 am

Would I have to do the experiment three times to make it a fair test?

If there is only one variable that needs to be changed, I have three different moulds, would that still count as one whole variable.

deleted-71447 · Post by **deleted-71447** » Wed Aug 23, 2006 8:03 am

Yes, you would need to run the experiment multiple times to answer a question about frequency ("how often. . .") .

Do you have three moulds of one fingerprint, or three moulds of three different fingerprints?

deleted-2131 · Post by **deleted-2131** » Wed Aug 23, 2006 10:17 am

In general, the more times that you run an experiment the better. I would reccomend running at least 10-20 tests with each imprint; more if you have the time. The more data you have, the better your project will be!

helper · Post by **helper** » Fri Aug 25, 2006 3:26 am

I am doing three moulds of one fingerprint.

Now there is a problem. I went to a biometric company today, to use their fingerprint scanner and tested my question if it worked or not, the gelatin that was used to create the artificial fingers melted before i even got there. Even if it didn't, it was still too soft for the fingerprint scanner to read.

Now i can't do it again, because i don't think i can go back to that company again since it would do the exact same thing again and is very far away from where i live. So is there another possible way to prove my question (guess i have to reword it) right or wrong? And must be in short time...

One of my ideas are to put my finger on some sort of a ink stamp then place it on a piece of paper to see the fingerprint. Then do the same thing with the artificial fingers and I see if they are similar or not...

But thats just an idea, is there any other ways I can do this in little time?

deleted-71447 · Post by **deleted-71447** » Fri Aug 25, 2006 12:18 pm

Sorry to hear that you ran into difficulties with your first attempt. Yes, you could make fingerprints on paper, and then ask friends to try to identify which fingerprints are original and which prints are fakes. Alternately, you can download trial versions of fingerprint recognition software and see how they perform with scanned image files of your fingerprints. For example:
http://www.neurotechnologija.com/download.html

It would be really interesting to see how the computer program performs compared to human subjects.

deleted-71447 · Post by **deleted-71447** » Fri Aug 25, 2006 12:27 pm

"You could make fingerprints on paper, and then ask friends to try to identify which fingerprints are original and which prints are fakes"

I think I was a little unclear with that suggestion. An appropriate procedure would be to show your friends an original print (made with your finger) and then an unidentified print, and to ask them if the unidentified print was made with the same finger, or with a mould. Of course, you'd want to make many repititions of the same procedure using different unidentified prints.

helper · Post by **helper** » Fri Aug 25, 2006 4:01 pm

Thanks for the idea. For the unidentified prints, do you mean different fingerprints or the same fingerprint but with different moulds?

Also, could you tell me what is the independent and dependent variables in this experiment?

If I download the fingerprint software, how would I be able to use it if I don't have a fingerprint scanner?

helper · Post by **helper** » Sat Aug 26, 2006 1:24 am

I don't think it can still be "how often can you fool a fingerprint scanner?" so what do you think the new question should be for this experiment?
Also, I don't think there is anthing wrong with this question "how hard is it to fake a fingerprint?"

helper · Post by **helper** » Sat Aug 26, 2006 1:50 am

One more thing, would I have to put down my previous method which failed (the one with the fingerprint scanner), on the display board or not?

helper · Post by **helper** » Sat Aug 26, 2006 2:46 am

Terribly sorry for another post. I don't think i worded my previous post right. 'Could' I be able to include the previous experiment and the new experiment I am doing now? Because the person who helped me out using the scanner in the biometric company said if the gelatin was a little harder, it would have probably fooled the fingerprint scanner because it had all the ridges, valleys, etc of a regular fingerprint.

deleted-2131 · Post by **deleted-2131** » Sat Aug 26, 2006 12:23 pm

Yes, I would definately include the the part of your project that "failed." It shows the judges how hard you worked on your project and shows them that you really do know how to think like a scientist--they didn't land on the Moon on the first try!

The only problem with the original question "How hard is it to fake a finerprint?" is that it is not very measureable. (e.g. the answer to this question could be "easy" or "hard") A science fair question needs to be measurable. Maybe you could change your question to say "How often do people mistake fake fingerprints for real ones?" You are still answering the same basic question, but now that question is measurable because the answer can be given in a percentage.

If you have any questions (or are now thouroughly confused) don't be afraid to ask!

helper · Post by **helper** » Sat Aug 26, 2006 4:00 pm

Thanks, is the dependent variable the one thing you have to change and the independent variable is all the things you keep the same in the experiment?

deleted-2131 · Post by **deleted-2131** » Sat Aug 26, 2006 7:40 pm

Very close. The independant variable is the thing that you change; the dependant variable(s) are the things that change in response to the independant variable. Controls are the things that you keep the same in the experiment.

For example, for your project, you are changing the type of fingerprint (real or fake). This is the independant variable. The answer of the person (they think the fingerprint is real or they think the fingerprint is fake) is the dependant variable. Some controls might be the size of the prints, the color of paper the prints are made on, etc.

Thanks for asking, because this topic (variables and controls) confuse lots of people!

Fingerprinting

Fingerprinting

Variables and Controls