By quickly looking at the algorithm it seems more suited to image/video processing.
The problem with such image processing based techniques is that they are highly analytical, and they are more likely to leave some artifacts that might not be suitable for games due to the dynamic nature of the games, unlike movies or images were the data is previously known. This is even mentioned in the article, which makes it not so suitable for games. Quoting from Wikipedia:
- The algorithm may need user-provided information to reduce errors. This can consist of painting the regions which are to be preserved.
With human faces it is possible to use face detection.
- Sometimes the algorithm, by removing a low energy seam, may end up inadvertently creating a seam of higher energy. The solution to this
is to simulate a removal of a seam, and then check the energy delta to
see if the energy increases. If it does, prefer other seams
Other reasons not to use such approach is that it's more likely to be expensive performance wise, but this needs to be proved by actually testing it.
Keep in mind that games use much different techniques to output for multiple resolutions that usually don't use traditional image processing techniques, as they can usually natively render to a render target (or multiple) with the target resolution. Think of traditional image processing techniques as the ones require analyzing pixels in color domain or frequency domain. I said usually because some Anti Aliasing aliasing techniques do some analysis on the image. Even though I don't know how widely they are used in games.