Raspberry Pi 5 - How fast is OpenCV Face detection?

A simple but extraordinary approach to explaining a fairly complicated subject.

awesome project, but I cannot get it running on a PI5 using bookworm 32 bit- what OS did you use on your PI5 ?

this command works sudo apt-get install libopencv-dev python3-opencv
the next command pip3 install cvzone has to be run within a virtual environemnt so
python -m venv /home/pi/gesture
cd gesture
source bin/activate
pip3 install cvzone
it eventually fails with an error message
" with open(config_py, 'w') as opencv_init_config:
^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: '_skbuild/linux-aarch64-3.11/cmake-install/python/cv2/config-3.py'
[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for opencv-python
Successfully built numpy
Failed to build opencv-python
ERROR: Could not build wheels for opencv-python, which is required to install pyproject.toml-based projects"


Using the 64 bit bookworm it installs, the camera works using libcamera-hello but the face recognition fails with an error along the lines of cannot get image.

can you use it to detect spicific face and connect to bt device/turn off bt?

Thanks for sharing. You are very talented.

Awesome, well explained. I have this up and running on a Pi4 it should be smoother on my Pi5 when it arrives. More ideas for projects come flooding into my mind. Keep up 
the good work.

What is the approximate detection frame rate?

Impressive project/speed 🙂

Could you try to show fps of tensorflow lite and yolo on pi5 ?

Great video! Love your fan girl!

Surely it should detect you as a Java enthusiast?

Good video. btw I had exactly the same problem with some eSun PLA+