[sp332]

You need a neural network that can understand the content and the style separately. It needs to interpret the content out of the frame and then make a new frame with that content but in a new style.