| title | Thinking in Blender: Staged Executable Inverse Graphics with Vision-Language Models |
|---|---|
| authors | Guangzhao He, Rundong Luo, Wei-Chiu Ma, Hadar Averbuch-Elor |
| date | 2026-06-01 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2606.02580 |
| upvotes | 3 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Inverse graphics is a longstanding and highly underconstrained problem that seeks to reconstruct images as editable 3D scenes which can be rendered, relit, and manipulated. In this work, we investigate whether pretrained vision-language models (VLMs) can perform executable inverse graphics directly from a single image by reconstructing a scene as an editable Blender program, without relying on specialized 2D or 3D foundation models, differentiable rendering, or multi-view supervision. We introduce Staged Executable Inverse Graphics (SEIG), an agentic framework that reconstructs a 3D scene from a single image by progressively refining scene factors including geometry, materials, composition, and lighting directly in executable Blender code space. We evaluate our framework across diverse scenes using a range of reconstruction metrics spanning pixel-level, perceptual, and semantic fidelity. Our experiments show that staged reconstruction substantially improves reconstruction fidelity, highlighting the importance of task decomposition for executable inverse graphics with general-purpose VLMs. Finally, we showcase various downstream applications enabled by the reconstructed editable Blender scenes. |
cite
@misc{he2026thinking,
title={Thinking in Blender: Staged Executable Inverse Graphics with Vision-Language Models},
author={Guangzhao He and Rundong Luo and Wei-Chiu Ma and Hadar Averbuch-Elor},
year={2026},
eprint={2606.02580},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2606.02580}
}