Want to protect your cyber security and still get fast solutions? Ask a secure question today.Go Premium

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 592
  • Last Modified:

Urgent: Implementing Face Recognition System

Dear Experts,
I have a problem regarding face Recognition system. I want to design a face recognition system using neural networks in matlab. I'm trying to implement Eigenfaces Algorithm but I don't know how. Can anyone  please tell me how to implement it? or give me any other algorithm to recognize gray scal faces.

Thanks alot for your help.
0
Al-Qershi
Asked:
Al-Qershi
  • 3
  • 2
1 Solution
 
InteractiveMindCommented:
Or, you could just fail like so many other Scientists who have attempted this...

This is a huge project, and although I'm sure you realise that it won't be easy, I don't think that you truly recognise how difficult this will be.

Teams of experienced scientists have failed this sort of project; there's also teams working for the CIA at the moment, working on face recognition systems, but have been working on them for years, but haven't yet succeeded.

However, you may be a genious... or just lucky ;)  So, if you really want to do this, but don't want to join one of the many, free, online projects (where teams of people all over the globe are working on this sort of thing), then perhaps some of these links will help a bit:

  Matlab Face Recogition System: http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=4408
 
  Information on the Eigenface Algorithm: http://pubweb.northwestern.edu/~acb206/ece432/FaceRecReport.html

  And you may be interested in these:  http://www.accusoft.com/imaging/visiquest/home.asp
                                                        http://www.attrasoft.com/

Best of luck with it,
>> IM
0
 
Al-QershiAuthor Commented:
Thanks alot but I think i will give it a try
0
 
InteractiveMindCommented:
Sure. Those links cover what you need to know really.
I once developed a similar system in Java; but I rarely speak of it, because people assume that you simply place the photo in, and let it get to work -- but this is wrong. What my attempt did, was allowed you to specify approximately the angle that their head was at, then you had to draw lines between certain places on the head - so that it would calculate the distance ratio between the eyes and the height of the head, etc.
It then used these ratios to recognise certain people, and worked most times, even if the person in your existing database was facing sideways, and the face you put in was facing forward; it would still be very accurate (although I only inputted a small amount of photos, etc.)... to allow this method to be consistent even with photos at totally different angles, you have to use quite a lot of Maths (but, mostly stuff like vectors, pythagoras, trig, sine/cosine/arccos, etc).
Definitely something I'd consider doing again one day -- but just don't set your heart on developing something that might be used by the Feds - lol  :)

Enjoy.

>> IM

PS: Are you doing this because you're interested in it, or is it some Uni project?
0
 
Al-QershiAuthor Commented:
Thanks IM for your help but it is much easier in matlab. so far i recognize almost 70% of the testing data base. It is not great, but good enough.
Actually I'm doing this for the Feds....................... just kidding :D.
I'm doing this for a project in my university. I'm taking a course called " Humans machines interactions". it is a little bit hard, but it is very intersting.

Have fun

AL-Qershi

0
 
InteractiveMindCommented:
Fantastic. Best of luck with it   -->  And thanks for the Grade 'A'  :)
>> IM
0

Featured Post

VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

  • 3
  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now