It’s been quite a while. My query now concerns the rationalization of image-to-3D-converted geometry.
My current process involves converting AI-generated images into a base 3D model.
This is followed by mesh-to-rationalization, where I refine and optimize the model within Rhino.
I’ve experimented with splitting the mesh using Brep and remeshing with QuadRemesh.
From here, I plan to rebuild parts of the structure, exploring methods inspired by Frank Gehry’s physical model-to-digital rationalization process. I also plan to develop a grid shell for this once I recreate the entire geometry.
I’ve also been reading many relevant posts on Discourse from Daniel Piker, Laurent Delrieu, Anders Deleuran, and many other members (apologies for not remembering everyone’s name).
However, I’m still open to any good ideas. My main goal is to effectively use image-to-3D-generated meshes.
If anyone has insights or alternative approaches for complex mesh rationalization, I’d love to hear your thoughts!
Like how to recreate this oval facade structure like this in the attached image so that i can detail it out the core structure and wooden batten panels. But I need to with image-to-3d mesh. Kind of reverse engineering with the 3D mesh.
Ahhh, in that case you might implement QuadRemesh (or some other remeshing method/plugin that yields uniform quads). Followed by a dynamic relaxation routine (i.e. using Kangaroo2) optimising the mesh towards a torsion free offset geometry with planar faces. Daniel has posted extensively about this topic, see e.g.:
Could you please shed some light on this? Despite multiple attempts, I’m still struggling to achieve my intended outcome with the 3D mesh model I have. I understand that I could remodel the entire form, but that would make the generated 3D mesh unusable, which I’d prefer to avoid.
