In the ChromaGAN paper published in WACV 2020 by Patricia Vitoria and Lara Raad and Coloma Ballester, we saw that they pointed out that "ChromaGAN shows variability by colorizing differently some objects belonging to the same category that may have several real colors". While it could be one of its strengths, it is actually a downside for colorizing videos. In a video, some images belong to the same class, and so may have different colorations. We applied ChromaGAN on a short surfing video, and here are the results:
We can see that the colorized video hesitates between a blue-ish and a more yellow-ish coloration.
Here are all the colorized frames: