Web based face recognition system

I am thinking of building a web based face recognition system. I know there are a few li开发者_开发百科ke KeyLemon, and others offered by different manufacturers that allows the laptops users to login into Windows using their face. I am wondering if this functionality could be transfered to a web application.

suggest you use this as the basis

OpenCV (Open Source Computer Vision) is a library of programming functions for real time computer vision.

There was an excellent podcast on OpenCV on Hacker Medley which has various references that are useful. From that i understand that the library tends to move quite fast in development terms so needs close attention.

You may use thing like flash to access to camera ... , and then use the same algorithm to recognize the face ..

I've written a web application which does something similar. And I have to say - I'm quite disappointed at the level of technology we're currently at for such things. The system in question used a 10mpix Canon camera and a special flashlight stand. It had to have a perfectly white background, the head had to be tilted exactly the right way, couldn't be rotated by more than a few degrees, and had to have very precise distances to the edges of the picture. And even then it gave a lot of false positives and negatives.

So maybe they've come up with something better today, but I doubt that. This was all 2 years ago and the software was some commercial product by a company that specializes in that sort of thing.

So all in all I say - better don't. Biometrics are cool, but currently they are way too unstable to be deployed in anything more than niche situations.

Keylemon provides web api to enroll faces and their later recognition. You can use this web apis to integrate in your application to provide face recognition functionality. It works like this. During enrollment six photographs are taken and a bio metric model is generated. A model id is returned to the client. This model id needs to be stored in application database. For face recognition, web camera streaming combined with model id is passed to the keylemon server. If the model id and stream matches the face is authenticated.