Author |
Message |
Steve Everist
| Posted on Sunday, November 09, 2003 - 09:29 pm: |
|
It is so incredibly simplified that it really doesn't hold any weight good or bad. Not only does it only include level 2 detail, but it doesn't distinguish between the different level 2 detail. Nothing at all is mentioned about level 1 detail either, let alone extending to level 3. This model completely ignores everything that is involved in fingerprint identification. ...oh, if the statistics were so easy to model! |
Kasey Wertheim
| Posted on Saturday, November 08, 2003 - 10:25 am: |
|
David, You are right... I misread his statement. But still, as you mention, a very simplified model. And the overall point is that this is just level 2 detail. Add level 3, and statistical probabilities offer even more support for the concept of biological uniqueness. -Kasey |
David Fairhurst
| Posted on Friday, November 07, 2003 - 09:07 am: |
|
Not quite Kasey. K.S. model says 1/100 to the power 10 that's 1/100000000000000000000 Smaller that 1/1000 but still far too big a number because of the very simple model that he is applying. |
Kasey Wertheim
| Posted on Thursday, November 06, 2003 - 10:11 pm: |
|
K.S., To expand on Michele's post, I agree that your analysis is not complete. 1/100 times 10 is 1/1000. By your mathematics, every 1000 fingers would be the same, or one out of every 100 people would have the same fingerprint. The fact is that no statistical model just takes into account a feature at a location. There are different features with different directionality. And furthermore, there are different positions within the fingerprint (related to the core) and different features have different shapes. Currently, no computer model accurately takes into account ridge shapes, pore positions, etc. that examiners compare. Therefore no computer model approaches the accuracy of fingerprint identification. However, current statistical models support and uphold the biological uniqueness of friction ridge skin to numbers thousands of times the entire population of the earth. So even though mathematics cannot prove biological uniqueness, it still supports uniqueness. And it doesn't even have all the information available that examiners do. It is a great science that we are a part of. |
Michele Triplett
| Posted on Thursday, November 06, 2003 - 09:06 pm: |
|
K.S., This is not at all accurate. If it were that simple, then it would be easy for statisticians to come up with a number of characteristics that would always establish individuality. From http://www.latent-prints.com/iai_standardization_committee.htm "In 1973, The IAI Standardization Committee released the results of a three-year study. They recommended and adopted that "no valid basis exists at this time for requiring that a pre-determined minimum number of friction ridge characteristics must be present in two impressions in order to establish positive identification." This was based on the fact that each print has a unique set of circumstances." Michele |
K. S. Anil
| Posted on Thursday, November 06, 2003 - 02:30 pm: |
|
An average fingerprint contain atleast 100 charecteristic points.It can be assumed that the probability of occurence of a particular point in a particular position in a print is 1/100.So probability of a point in a print repeating in another print is 1/100.Then the probability of repetition of 10 points of a print in another print simultaneously is 1/100 to the power 10. This is a number very well equel to zero. So fingerprint never duplicates. |
Webmaster
| Posted on Sunday, May 21, 2000 - 09:11 pm: |
|
Jane, That's a very open-ended question. Sometimes we point out that the science of friction ridge identification lends itself well to mathematical verification of the fundamental principles of permanence and uniqueness through AFIS (automated fingerprint identification systems). Indeed, in at least four states in America, inked fingerprints are daily processed by the thousands with no human intervention insofar as analysis, comparison and evaluation of the identifications against existing fingerprint records.... with a zero error rate. Since computers are in essence math systems, this is one way in which mathematics are related to fingerprinting. --Ed German |
Jane Scharmer
| Posted on Sunday, May 21, 2000 - 09:10 pm: |
|
What does mathematics have to do with fingerprinting? |
|