I use Keil MDK 5.20. I have 2 boards.
STM32F4DISCOVERY - 32-bit Cortex®-M4F STM32F407VGT6 MCU, 1MB Flash, 192KB RAM (Max freq. 168 MHz) 32F746GDISCOVERY - 32-bit Cortex®-M7 STM32F746NGH6 with FPU, 1MB Flash, 340KB RAM (Max freq. 216 MHz)
With the 32F746GDISCOVERY, a connector P1 with the DCMI(Digital Camera Interface) signals is available to connect a camera module such as STM32F4DIS-CAM (ST order code).
Since both boards have example codes for getting data from the camera module,
I tried to plan to use OpenCV for face detection with those boards.
The latest versions of OpenCV is 2.4.13 (2016-05-19) and 3.1 (2015-12-21).
I want to use those libraries to detect faces with the DISCO boards.
Questions: How can I compile and build the OpenCV libraries using Keil MDK 5.20?
-Best Regards, David
http://opencv.org/
answers.opencv.org/.../
Looking at the supported platforms - opencv.org/platforms.html - I don't think your boards are in the same league?
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?
There will not be any other libraries that will require less processor capacity.
You will just need to spend the time trying to compile whatever code you have access to, and then evaluate exactly how much processor resources it does take to evaluate a single frame.
If analyzing a still image, it might be enough to spend a number of seconds. If handling video, then the requirement might be real time analysis in which case you only have tents of milliseconds for the processing.
So just experiment and see what comes out - if the processor is too small, then you have still learned something and can later go get a more powerful processor board and try again.
Maybe see this first, shervinemami.info/embeddedVision.html
And this. charmedlabs.com/.../