Select object on video without AI model

[crackwitz]

then get any one of the trained models that are out there