Just made a recap from a project which was laying around for quite some time.
The idea is very basic but the outcome is always different depending on the audience.
I do a recording of people in a box who turn a black and white sheet on specific timings.
Al these recordings or placed in a grid making it a human operated pixel display.
Did you create small image sequences from video? And did you use OpenCV or similar to automate the process and detect the white and black images or did you fully rely on the kids timing?
When someone is in the box they see themselves on the screen where they follow the basic instructions on when to turn the black and white sheet.
Most of the time other people are shouting black, white
And as with every screen we have “broken” pixels. But all this creates super nice variation and randomness.