PCA for authentication.

Hello all, I have used PCA for recognition successfully. I was wondering if we can use it for Authentication. What I mean, I want to get a degree of match of the test sample with all the existing training classes. like 79% for class 0, 60 % for class 2. At present it give output as class that has highest match. I want to get matching percentage with all the training classes.

Thanks, -Piyush

answered Mar 27 '14

GilLevi

1357 ●10 ●27 http://gilscvblog.com/

You can get the confidence level for the predicted class:

Ptr<FaceRecognizer> model = createEigenFaceRecognizer(10,2000000000);
int predictedLabel = -1;
double confidence = 0.0;
model->predict(face, predictedLabel, confidence);

Comments

Have you understood the question ? I want the confidence of test image with all the training class.

piyush (Mar 30 '14)edit

I don't think you can extract that without changing the face recognizer API. The confidence level is the closest thing you can extract.

GilLevi (Mar 30 '14)edit

We can tweak into the code, are all the codes available to us for changing, or incorporated inside object files .lib we use in windows. If you can guide on this.

piyush (Apr 2 '14)edit

I'm sorry, but I'm not that familiar with the code. However, I would be glad to take a look and try to help you with that, but I'm afraid I would only get to it in a few days. You can mail me at gil.levi100@gmail.com so we'll continue discussing there.

GilLevi (Apr 3 '14)edit

add a comment

PCA for authentication.

1 answer

Comments

Links

Question Tools

Stats

Related questions

PCA for authentication. edit savecancel

1 answer

Comments

Links

Question Tools

Stats

Related questions

PCA for authentication.