That’s a common technique to try and fool copyright detection algorithms: image manipulation (by flipping the footage) or audio manipulation (by shifting the pitch or tempo, as an example).
I don’t believe a copyright detection algorithm should have any problem identifying a flipped image.