ISF [Iriscan Fingerprint Identification]

[vinaya]

dear all
for a biology project
me and another girl are working on combining a iriscan and a fingerprint identification system to make a absolute foolproof system
in case of iriscan, people with drooping eyelids, long eyelashes etc find it hard to use.in case of fp, there is always the chance of skin disease,injury,etc.we are planning to use our system at a global level and these anomalies come in the way.
so we planned to combine both to produce a new software.do you guys think its a good idea? can u think of a good idea to actually build a working model for our project?
or do you have any doubts?
please do post.thanks

[mith]

Both technologies already exist...how are you combining them?

[vinaya]

actually the technicalities are not something we are clear about after all
we are highschool students.this is really where we require help.
however we are damn sure that the idea can work as the software is possible.thanks so much

[mith]

So do you need a working prototype? Or just a software package? Iris scans are less common but fingerprint technology is available for laptop computers. I don't really see how you actually will do anything other than getting the response from each of the equipment.

[vinaya]

we think we can get hold of the software as they do give free demos for a month or so
its the working prototype that we are worried about

[MrMistery]

Why don't you use retin scan? Retin scan is absolutely sure and it is impossible to cheat it. Or that is what they want you to think...

[vinaya]

the thing about ret scan is that the retina can be easily damaged and it does show a few changes during the course of years thats the prob with it

[vinaya]

we are using dual id as we get a clearer, larger database
The main reason is, it gives a wider pool of characteristics to choose from. Let us say, we have five distinguishing characteristics in iris identification – its diameter, colour, texture, spots and so on. The first two would be easily obtained by the computer and the PARE software to translate this would be easy, but the other three will be more difficult to handle. The problem is unless you have more characteristics to compare, the chances of error will be moreFor example if you have just the colour and size of the iris, the identification will be correct to 1 in 10000. That is to say, if you have a population of say 20 lakhs in chennai, you will have 200 people shortlisted by the computer with just these two parameters.
To improve the accuracy we need more parameters….
This is where the dual id concept comes in… There are easy pickings in the finger printing.. like the whorl or disc formation of the lines, number of concentric circular pattern etc. If two of these are added to the computer, let us say this has an accuracy of 1 in 10000 on their own, the accuracy of the dual system would be 1 in 10000 times 1 in 10000, that is a mind boggling 1 in 10 crores. (The figures given are just to illustrate the concept, they are approximate)