4 ms·
> I wonder if there's a way to describe and interactively refine descriptions of areas of interest in a symbolic way—as Lisp trees if not Unix strings—maybe usi
by investinme 7y ago
> I wonder if there's a way to describe and interactively refine descriptions of areas of interest in a symbolic way—as Lisp trees if not Unix strings—maybe using existing neural networks. Certainly a speculative idea, of course.
I think the future is more like saying to the computer what you actually want. Like "delete the apple from the picture".
Photoshop already has a primitive form of this particular action, called Content Aware Fill, where you draw a rough selection around an object you want to delete, and it will fill a plausible background in its place. Works pretty well.
https://helpx.adobe.com/content/dam/help/en/photoshop/using/content-aware-fill/jcr_content/main-pars/procedure/proc_par/step_1/step_par/image/Launch-the-Content-Aware-Fill-workspace_2.png https://helpx.adobe.com/content/dam/help/en/photoshop/using/...
- kragen 7y agoIn a sense these are different uses of a computer for graphics: one is a medium of human visual expression, which is Photoshop’s forte, while “delete the apple” is more like delegating a functional task to an independent contractor. I don't think one is likely to entirely replace the other, just as Photoshop hasn't replaced netpbm, photography hasn't replaced painting (even if we do more painting in PS than with paint nowadays), and the city bus hasn't replaced dancing or skateboarding, even though all of them are ways to move your body. There are certainly some uses of PS that things like caption-to-image GANs will replace. Analogously, last night I spent quite a while with the expressive, high-bandwidth, direct-manipulation interface of a box cutter cutting cardboard, for something I'd much rather be doing with a laser cutter cutting shapes generated from a constrained-optimization algorithm. At least this time I managed to only cut the cardboard. Texture synthesis is a really interesting tool for both purposes. There are a lot of amazing papers in recent SIGGRAPHs with new algorithms for this kind of thing.