Thanks for your reply Andrew.
Then, let me change my question.
For these Cortex-M processors, is there an alternative libraries (instead of OpenCV) for face recognition
which can be compiled/build with Keil-MDK 5?
-Best Regards, David
Here is a starting list: facedetection.com/.../ includes a list of general face detection and/or recognition libraries (in the bottom half). See === below for the libraries that qualify.
I think we have to point out that face detection and face recognition are two zdifferent animals, but face recognition is the end goal.
I haven't tried any of these out, but I guess a fast Cortex-M7 would be the most capable to handle the high video-processing demands, and users require the C/C++ source code either to use directly in the application or to rebuild the library with the ARM compiler (because the library was probably created using another compiler in the first place).
===dlib C++: blog.dlib.net/.../dlib-186-released-make-your-own-object.html
===CCV: http://libccv.org/ This is a C library under a BSD 3-clause license.
===Associative Memory for Face Memorization and Recognition: Free fully self-contained C++ code for on-the-fly storing and retrieving of patterns, such as binary coded faces, using projective associative memory. (by Dmitry Gorodnichy / Computational Video Group / NRC-CNRC) videorecognition.com/.../ (Thank goodness for the Internet Wayback Machine...)
The first link goes on to mention commercial libraries, and I think the third library offers licenses for the author's more advanced libraries.
Interesting, something called the "ARM Compute Library" is going to become available: www.electronics-eetimes.com/.../1
It will be open source, and it will contain some machine learning and computer vision capabilities. Let's see how much applies to the Cortex-M world vs. Cortex-A with Mali GPU...