AI Paper Atlas

Research map / 3D vision / Face and shape

Efficient Autoregressive Shape Generation via Octree-Based Adaptive Tokenization

Kangle Deng, Hsueh-Ti Derek Liu, Yiheng Zhu, Xiaoxia Sun, Chong Shang, Kiran S. Bhat, Deva Ramanan, Jun-Yan Zhu, Maneesh Agrawala, Tinghui Zhou

ICCV 2025 · 3D vision › Face and shape · approach: 3D shape generative modeling · 9 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add quadric-error adaptive subdivision to the octree tokenizer (ablation)OAT (discrete) · Objaverse testIoU: 84.2 -> 86.7Improved
add adaptive subdivision at similar token count (607 vs 625 tokens) (ablation)OAT (discrete) · Objaverse testIoU: 88.3 -> 89.7Improved
add adaptive subdivision at similar token count (1726 vs 1284 tokens) (ablation)OAT (discrete) · Objaverse testIoU: 89.9 -> 90.2Improved
add adaptive subdivision for continuous latents (607 vs 625 latents) (ablation)OAT-KL · Objaverse testIoU: 92.1 -> 92.7Improved
use OAT tokenization instead of fixed-size Craftsman-VQ at about half the tokensOAT (discrete) · Objaverse testIoU: 83.8 -> 88.6Improved
increase OAT token count from 266 to 1284 (ablation)OAT (discrete) · Objaverse testIoU: 86.7 -> 90.2Improved
replace Craftsman-VQ+GPT with OctreeGPT on OAT tokenstext-to-shape generation · ObjaverseFID: 85.10 -> 56.88Improved
use breadth-first instead of depth-first node ordering (ablation)OctreeGPT · ObjaverseFID: 67.31 -> 56.88Improved
use breadth-first next-token instead of next scale prediction (ablation)OctreeGPT · ObjaverseFID: 198.37 -> 56.88Improved
add hierarchical octree structure over leaf-only tokens (ablation)OAT · Objaverse generationKID: 13.65 -> 5.79Improved
use convolution pooling instead of average pooling in encoder (ablation)OAT encoder · ObjaverseIoU: 88.5 -> 88.6No clear effect
use concat + MLP pooling instead of average pooling (ablation)OAT encoder · ObjaverseIoU: 88.6 -> 88.6No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy