📢 GenRecon Code Release 📢
Few images in → complete, high-fidelity 3D scene out!
GenRecon builds a generative prior on full scenes, resulting in unprecedented 3D reconstruction quality.
🔗 https://github.com/kasothaphie/GenRecon 🌐 https://kasothaphie.github.io/ 📄 https://arxiv.org/abs/2605.23888
Users praise GenRecon's code release for extending Trellis2 as a generative prior to produce pixel-aligned 3D scenes from images.
No Digg Deeper questions have been answered for this story yet.
Core idea is to leverage Trellis2 as a generative prior, which we:
a) expand to entire 3D scenes b) make it pixel aligned to match input images c) generalize to multiple input views for large scenes
Great work by @katha_schmid, @nicolasvluetzow, Jozef Hladký, and @angelaqdai.
📢 GenRecon Code Release 📢
Few images in → complete, high-fidelity 3D scene out!
GenRecon builds a generative prior on full scenes, resulting in unprecedented 3D reconstruction quality.
🔗 https://github.com/kasothaphie/GenRecon 🌐 https://kasothaphie.github.io/ 📄 https://arxiv.org/abs/2605.23888
