I am creating a fuzzy set for a biometric signal. The case is when a user places his/ her finger on a fingerprint reader, obviously some pressure will be applied. If s/he applies VERY LESS/ light pressure on the scanner, there are chances that finger minutiae will not be properly extracted; now if a person applies MORE/ TOO FIRM pressure, enough noise can be generated that might not necessarily be removed via SNR technique.
I am searching the values in Pascals (or any other unit) for fingerprint when a user apply finger pressure.
I am searching the following values to fill
- Too Firmly = ___ Pascals (Pa)
- Very Firmly = ___ Pa
- Properly = ___ Pa
- Marginally less = ___ Pa
- Very less = ___ Pa
You might try putting a pressure sensor under the fingerprint scanner, and reading the pressure off of that. To get the numbers you want, then you read the Pascals of the pressure, and study the quality of the fingerprint that was scanned. From the results you can see what your values are for your matrix.
Hi Yousaf,
I am not related to these kind of measurements (pressure measurements), but I guess you can simulate it with an analytical balance, which calculates weight by converting the strenght applied to mass with "kilopondios" (kp, pressure which is applied with an object of 1 kg with normal conditions of gravity).
I think that you may simulate your different cases (too firmly, firmly, properly, and so on) in such instrument.
Hope it helps, but it might be too simple or not enough reliable/accurate for the final aplication.
maybe width of edges can help. wen we press our finger too firm, edges are bolder. otherwise thinner. but its not if and only if !
I agree with Faezeh. You can probably use the total number of extracted pixels as a relative gauge of pressure. For example, if someone presses too lightly, only a light image with thin lines will appear. If someone is pressing hard, the lines will become fatter, and the overall image will become darker.
If you are taking a gray-scale image from the scanner rather than black and white, you could compute a sum of the darkness (overall intensity) for each pixel and use the sum to rate individual fingerprint scans.
Dear Yousaf
the best solution of this problem you can collect many images (ex, 10 images) for each person and then apply fusion for that data to generate a new features which are the reference image.
Regards
Thanks guyz, your ideas ultimately lead me to make an experiment on weight machine that is typically used in markets. 30 gms for normal finger pressure was experienced and converted finally in Pascal that is equivalent to 0.29 Pa.
You will find that fingermarks deposited with increased pressure will exhibit wider ridges and thinner furrows. You should also consider ridge pattern distortion as a result of an uneven pressure applied during deposition (the extent of which increases with deposition angle and pressure). As suggested previously, the use of a balance during a series of controlled depositions would be optimal to determine the values you are searching for.
All the best for your research.
- Badges
- Science topic