🚨 Paper Alert Our recent breakthrough CAST: Component-Aligned 3D Scene Reconstruction from an RGB Image has been accepted by ACM SIGGRAPH 2025 Journal Track! CAST will change the way create scenes in 3D Art and Embody AI. 🚀Soon available at 👇Details

(1/5) ArXiv: Project Page:

(2/5) The input RGB image is processed through scene analysis to extract key information, followed by pose-aware generation to create initial 3D models. Physical constraint refinement ensures realistic interactions and spatial relationships, yielding a high-quality, mesh-based 3D scene.

(3/5) Network design of our alignment generation model (Sec. 4.2), occlusion-aware object generation model (Sec. 4.1), and an illustrative figure of the texture generation model.

(4/5) Comparison of scene reconstruction with and without relational graph constraints. By integrating relational graph constraints, our method ensures both physical plausibility and accurate alignment with the intended scene, maintaining correct spatial relationships.

(5/5) Qualitative comparisons of CAST with state-of-the-art single-image scene reconstruction methods. From left to right: Input image, CAST, ACDC, and Gen3DSR.

Sign up & create wholesome anime art on Moescape AI now!

@siggraph Waiting for proof with the original Myst game. If we can reconstruct those 3d scenes in high identity with techniques like this, that’ll open a lot of use cases for digital archaeology

@siggraph thats amazing, can't wait to play with it!