4 ms·
This is really cool, but the only thing I can think right now, the question that's eating away at my soul, is: why did it color-cycle the apples?
by rcthompson 10y ago
This is really cool, but the only thing I can think right now, the question that's eating away at my soul, is: why did it color-cycle the apples?
- isoprophlex 10y agoThere are also semantic masks applied (can be found in the github repo), I think that's how the Colorado swapping is done.
- rcthompson 10y agoAh, I see. They're right here: Input mask: https://github.com/luanfujun/deep-photo-styletransfer/blob/master/examples/segmentation/in23.png https://github.com/luanfujun/deep-photo-styletransfer/blob/m... Output mask: https://github.com/luanfujun/deep-photo-styletransfer/blob/master/examples/segmentation/tar23.png https://github.com/luanfujun/deep-photo-styletransfer/blob/m... So, presumably these masks were drawn manually. The three different colors in the input mask tell it to train 3 different models from the masked regions, and then the same three colors in the output mask tell it where to apply each trained model. So it trains a model for each apple and then applies that model to the next apple over. It's not just randomly swapping colors on the apples.
- alanbernstein 10y agoI'm afraid the answer, like the answers to so many other soul-eating questions, is that it's random.