Even if you’ve been living under a rock for the last five years, it’s difficult to escape the media hype surrounding artificial intelligence and how it will either elevate humanity to a transcendent state, or threaten to consign it to the pages of history.
Aside from teaching cars to drive themselves, we believe that a far more important application for this ground breaking technology has been found, namely using image processing and manipulation algorithms to clean up and upscale the graphics used in id Software’s seminal first person shooter, Doom!
Scores of forumites over at doomworld.com have been busy experimenting with various AI frameworks, including Nvidia and letsenhance.io, to upscale and infer additional detail into Adrian Carmack’s artwork that was never part of the original imagery, and the results have proven to be extremely impressive.
Admittedly, a reasonable amount of human intervention and manual touching up was required to achieve the various results in the thread, but creating and up-scaling the assets through purely manual means would have been nigh on impossible.
We know that many of our readers will prefer the pixelated imagery, as it’s part of the game’s iconic look, but there’s no denying that the use of AI in image processing and manipulation is incredibly cool!
If you’re interested in running the game using these up-scaled assets, it’s possible to download complete PK3 files to run the game with the updated artwork – be sure to take a look at the thread on Doom World for more details!