Hello, I’m a noob with a perplexing modeling challenge. I have a picture [down at the bottom] of an onion shaped dome home, and I’m trying to build an accurate model and cut the windows in the right place.
I traced the profile of the dome and revolved it, then used a spherical decal to project the image onto the model, but I can’t figure out how to stretch the image after I placed it so that it covers the entire front half of the dome.
Or is there a better approach someone might know about?
Hi -
You can use the FlowAlongSrf command to do this.
-wim
OK, I’ll have a look at that, thank you.
That particular A.I. picture represents a shape made out of multiple surfaces, complete with distinctive crease lines between the elements. If you want to recreate it properly in Rhino, you should build it from multiple surfaces, too.
AI slop.
A.I. kills the human imagination and desire to develop artistic skills. I have seen plenty of “car designers” all over the web offering paid services who actually have no idea how to build a simple cube in a CAD program. They just use Vizcom, Midjourney, Wix etc to generate a picture in seconds. In fact, several of those so-called “car designers” and companies who bought A.I. designs contacted me over the past couple of years to turn these into 3d CAD projects for them. I rejected all of them for obvious reasons.
Yes, I know, but I’m not trying to recreate the AI result, which is undoubtedly an incoherent mesh. I am using this as a training exercise for a client who shows me a random picture and asks how it can be built. So I need to proceed as if this was a real structure, and fill in what the AI did not.
Well, that’s encouraging for us who are committed to doing the actual work, but I think it’s important to acknowledge the growing presence of AI in workflows. For instance, do you appreciate some of the drudgery in making textures that AI alleviates, or are you a purist evangelizing against all forms of AI intrusion? Because it strikes me as odd that so many of the people who say that AI cannot ever be as competent as human designers are also saying that it should be banned. Both cannot be true at the same time. And it’s conceivable to me, as much as I wish it were not, that AI will eventually be able to use the same sort of judgment humans use to build more aesthetic models and also better 3D meshes than we can. It’s nice that they can’t do it yet, but let’s see what happens next year.
AI won’t take you job, the person using AI will, dismiss AI at your peril, whether you like it or not, the horse has bolted the stable, the train has left the station, its here and coming down the track, fast. Its like the second coming of the internet
As many already have said: “A.I. can’t replace me drinking beers” 
Persons who use A.I. to do their own job, as well as the robots, can’t replace the majority of professions. At least not in the next 10 generations. And definitely not at the same cost as using human workers. Companies who invest into robots and A.I. will seek for greater profits by eliminating as much workers as possible, however, that will NOT reduce the end price of their products but will increase their profits, eventually leading to consumers seeking for more affordable options like those offered by… humans. 
Not to mention that the population is becoming more stupid over the last two decades (thanks to social media, “smart” phones, intentional government politics etc) and the scenario of the film “Idiocracy” is very possible. The first few minutes of the film, and the last few minutes of the film, show exactly what’s happening now in the Western World…
You cannot get an accurate model by taking a flat image, wrapping it around a sphere and stretching it…
But why not ask AI to convert your AI image into a 3D mesh? There are several AI tools out there offering this service.
by design, of course
You’re right, that’s the quickest solution. When I started, I thought it would be easy to draw outlines on the 2D image and rotate the view to project them on the dome model and then slide them around until the 3D resembled the 2D. But it turns out to be much more of a hassle than I thought, and still looks weird–not a sustainable workflow for client projects.
For the actual sculpture at hand here (if I am understanding it) has the soft corners that align themselves to some subd type forms, but as a schematic approach to the image, taking that surface and projecting ovals onto the surface (work in plan to get the correct patterning) and then splitting that surface. From there offset/thicken that surface and start using the fillet and/or blend edge tools to get the rounded extents.
The AI is creating arbitrary forms (slop), there is some language there, but it has no rationality. You will have to re-interpret something to make it workable in Rhino as a coherant form. Unless you want to do something more abstract, in which case SubD - starting with a cylinder, playing with the profiles and so forth - would be your best bet.
Thank you.
