This is an experimentation of the OpenCV library to detect objects
My goal was to control my mousepad by gesture recognition through the webcamera, without using any external hardware.
In OpenCV, this is normally done by specifying a color range, and then by hoping that your webcam will detect the object of the corresponding color.
You can imagine that this is not the most robust way of doing gesture recognition. Since indeed the color detection was highly sensitive to external circumstances, I decided to implement things from scratch and do the same things with neural networks.
Some of the things that came out of this:
