Hi, I don’t know exactly how to formulate this question / request. I think I’m looking for something like a glossary of CG and A.I., and maybe somewhere where I can find programs and APIs based on them. Or a list of that too.
Well, I want to start studying things like photogrammetry for example, as we have here in the forum. But for this I first need to know the name of the technique. Only then look for some program that uses it, for example.
Other things would be:
That facial expression capture technique that came up in Avatar. There’s a really cool video of a woman on the internet made purely in CG. But I don’t even know the name of the technique.
Motion capture taken without markers and by ordinary cameras through AI also interests me.
Creature Evolve.
Machine learning used to create character animation. Including a technique I saw in a Siggraph YouTube video where a 3D human character, using motion capture or animation, adapted to the situation and environment based on these more advanced A.I., which used things like environmental physics, self balance, goals and paths, to make adjustments.
Another interesting video is of a wolf or dog walking around and adapting to the environment automatically based on need, apparently after using machine learning as well.
There’s a website that is capable of upscaling images with good quality, depending on the image and how much it would be stretched. Also using some newer AI. I don’t remember his name. Something that would be interesting to improve textures and a final render as well. Something similar has on another site, which can create frames among other frames improving the framerate quality even of an old video. I don’t know if it works in real time though. Anyway, it could also be used in game textures, for example. Automatic coloring is also interesting. And by putting it all together and a little more we would have a semi-automatic method to do something like remaster footage of the Great War, as it was done a while ago.
What about DeepArt sites that allow you to merge two images into one using the style concept? Loving Vincent could be done this way today. Not that this devalues the work of the film whose interesting part was that it was done entirely by hand-made paintings. But now think if something like this could be done in real time. Could a new redesign of games like Okami be considered?
And that new technique that allows the facial expressions or skin of a character in general to wrinkle and warp with more quality, and apparently in a more automated way than having to create textures that will be visible depending on a tension map.
I’m mainly a Blender user and although I know some techniques, terms and programs, I still feel that ‘things’ are missing if I decide to try to learn how to animate using something like what was presented in that Siggraph video along with things like unmarked video motion capture, that face animation method, and new softwares like Cascaduer.
I am not saying that I understand ML. I am still at the beginning of my programming studies. But being able to use techniques that involve them by tweaking some program, API, or something would be interesting to make me evolve as a digital artist.
I have to apologize for the size of the text and the difficulty of expressing myself. I believe that lists of techniques, methods, workflows, classes, programs, APIs, and new videos, so maybe a mix of glossaries, links, and tool names is what I’m looking for right now.
I really need this help. Thank you for your attention and patience.
