Getting familiar with ML as a design tool

I was thinking about what is “interesting” in the concept of movement. Or what does “interesting” even mean?

Some ideas I am thinking about to experiment with is whether the material will be able to pick up on the different parts of the body if we are not facing the camera. or if it is “smart” enough to know whether the different body parts in the frame belong to the same body. Although it might seem unimportant to know these points, we thought we should explore ideas like these to know the limits of the design tool.

We are still in the stage of playing around with the code to understand it, the sketch we are using now is the wrist distance. In the example below, the code was written in a way so that the camera would be able to pick up on more than one body. However, we managed to “trick” it.

Another observation we made was that the color contrast between the user’s clothes and the background is important. i.e. the less contrast in the colors, the more probable it will be for the key points and values to be jittery. The person needs to keep a certain distance from the camera for it to be able to pick up on the right body parts, which can be seen as a limitation. However, I think that this constraint could be viewed in a more positive way due to the fact that it implicitly forces us to work with the whole body, observe our whole body on the monitor while still in the stage of experimenting.

Regarding the coding part of the equation, are we going to be able to write the code in a way to support our idea? Does knowing in what level you are in programming influence you and your idea generating process from the beginning?