Metraipolis › archive › dc_bcaa756ef0d2
Octrees as an Explicit 3D Language — arXiv:2610.02388
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.02388
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T09:06:32.823Z
- SHA-256 of the archived text
6b8fdca4551e5267d3c533e8a062ddfaee34be12537b0c04de4c1ab1186f04b0- Stable URL of this page
https://metraipolis.com/documento/dc_bcaa756ef0d2
How to cite
Octrees as an Explicit 3D Language — arXiv:2610.02388 — arXiv (en), https://arxiv.org/abs/2610.02388 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bcaa756ef0d2 (Metraipolis archive dc_bcaa756ef0d2, sha256 6b8fdca4551e5267d3c533e8a062ddfaee34be12537b0c04de4c1ab1186f04b0)
BibTeX
@misc{dcbcaa756ef0d2,
title = {Octrees as an Explicit 3D Language — arXiv:2610.02388},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.02388},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bcaa756ef0d2; sha256 6b8fdca4551e5267d3c533e8a062ddfaee34be12537b0c04de4c1ab1186f04b0},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.02388. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Existing 3D large language models (LLMs) compromise on two fronts: they compress shapes into latent codebook indices or coordinate text, which removes spatial structure from what the model observes, and they acquire the 3D modality by fine-tuning the backbone, which overwrites its general language ability. We present OctLLM, which addresses both limitations. Geometry enters as an explicit 3D sequence of octree occupancy tokens. However, full octree sequences grow rapidly with depth; OctLLM therefore randomly empties penultimate-level nodes and omits descendants while preserving shape, yielding a shorter coordinate- and depth-anchored Sparse Octree (S-Octree) for position-aware mask-modeling generation and 3D understanding. On the other front, existing methods introduce a new modality with full fine-tuning or LoRA, but full fine-tuning is costly, LoRA limits 3D capacity, and both modify the language pathway. OctLLM instead adds 3D capacity in parameters separate from the pretrained ones: mesh tokens are routed through independent trainable branches in a subset of blocks while text and image tokens retain the frozen vision-language pathway, and the two streams interact through shared self-attention. It trains far fewer parameters than full fine-tuning, yet sets a new state of the art among unified multimodal LLMs, lowering image-to-3D FID by $17.4\%$ and raising render-grounded captioning by $28.7$ points over ShapeLLM-Omni, while matching the backbone on general language benchmarks. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-10-05 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.02388 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-06T08:21:44.139Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-10-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).