extracting landmarks for Facial recognition

[erangross27]

I have been building Algorithm in Java that extract face details, I am talking 68 landmarks of the face.
I want to use Rasberry Pi camera to run my code on it, the file + the machine learning model is around 100MB.
Is it possible to run it on the camera ?
I have no previous Raspberry Pi experience.

Thanks
Eran

[jwatte]

What file is 100 MB? What machine learning model are you using?

Yes, Raspberry Pi can run a variety of machine learning models, either using the CPU, or using the GPU if you have a version of it that can run on OpenGL ES.

(Note: Most "deep learning" machine learning packages aren't well supported on OpenGL ES, although I think caffe2 focuses on mobile and thus might be easier?)