Conclusions

Good

The app is lightweight, it doesn’t have any heavy algorithm, and as such, it is fast to calculate colors and detect the hand signal.

OpenCV and Numpy were the only external libraries of Python used, so it is easy to maintain.

The app is ready to use straight out of the box.

Bad

On the downside, there are some environments where the app doesn’t handle as well as expected.

Sometimes is hard to get a good color bound on the environment. If there is too much sunlight, it interferes as it isn’t as regular as artificial light.

As the signals are being calculated with areas, if the detection of the color is bad, the signals will have worse performance.

If the environment has colors that are matched with the skin color, there can occur that those areas merge with the hand and it makes impossible for the algorithm to calculate the signal.