There are two issues I'm skeptical about with this idea.
First, I'm not sure it performs well enough for games. It's probably a fairly performant idea, but I'm not sure it's fast enough for games. It might be but I wouldn't be too sure about that. Games are extremely performance-dependent.
Second, I'm not sure it's cost-efficient enough. I very much doubt it's possible to come up with a one-size-fits-all solution given the performance constaints, which means that there needs to be specialized solutions. In practice, this could easily mean one solution per game, i.e. training a neural network for each game that wants to use one. I'm not sure that's cost-efficient eough. It could be, because I'm mainly worried about and generating training data is as easy as just playing the game, but the overall development process probably still requires considerable effort.
Actually, what restricts this to just improve resolution? It's entirely possible a neural network could create graphical effects as well, for what is probably a comparable cost to that of simply improving resolution.
Existing User Log In
New User Registration
Register for a free account to gain full access to the VGChartz Network and join our thriving community.
