What algorithms are used in the Intel ® Perceptual Computing SDK 2013 Beta for the face detection and face recognition?
For example, In IPP Sample used Haar algorithm in face detection.
PCSDK 2013 beta uses proprietary face detection and recognition algorithms.
The face detection uses SURF-like features and the recogniton uses something similar to Gabor and SIFT features.
Thank for you comment
Did you do testing algorithms or technology of Intel ® Perceptual Computing SDK 2013 Beta by Face Vendor Recognition Test (FRVT) (NIST)?
Unfortunatly we do not perform face recognition testing using the FRVT.
Thank for your answer
Do you plan to spend any tests of face recognition?
I would like to know if you planned to improve the face detection/landmark algorithm for the SDK release.
Indeed i tried to use your functions in an implementation of 3D face tracking and it is not robust enough (lack of precision, latency, don't recognize face roll,...). To be honest i was quite desappointed with my tests in comparaison with other face tracking SDK using infrared projector/camera.
I think the perceptual computing camera would be an awesome tool for face/head tracking if it is using efficient face recognition algorithms.
Thank you for your answer.
Can you give some characteristics of face detection and face recognition algorithms in Perceptual Computing SDK 2013?
False detection rate
False Positive Error rate
False match rate
Average time for detection face
Average time for recoginion face
and so on
Thank you for youranswer