Face detection and comparison

[Sury@]

I want to know how can be Face detection and comparison can be done in a robust way.

I went through the other discussion forums and found that BiometricsView provides the tool kit for the face recognition. But I want to know methods that can be adapted for implementation. Where to start with.

[Shivam Awasthi]

Hi kpraveen,

Edge detection based face detection algorithms prove to be very robust ones. To make the algorithm more robust, do the thresholding of the face image on the local variances of the luminance data. This would take care of light intensity variations (which is generally a problem with these kind of applications). Generate the feature vectors and to classification. More the number of feature vectors used for training and classification, better the accuracy. Also people are now using 3D face detection techniques like AF3D and curvature analysis. You can try those. Also i dont think BiometricsView is providing solutions anymore. But you need to confirm that.

Hope this helps.

Regards,

Shivam

[mostafanfs]

Does anybody have a sample for face detection ? (NOT face recognition)

Thanks all