Alternatively, if you need a generic placeholder:
Training a model to find a card within a camera frame. midv260 link
Why it matters