AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Fast Training of Diffusion Transformer with Extreme Masking for 3D Point Clouds Generation

Shentong Mo, Enze Xie, Yue Wu, Junsong Chen, Matthias Niessner, Zhenguo Li

ECCV 2024 · Point clouds › Shape and surface completion · approach: Point cloud diffusion generation · 11 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 3D voxel masking (ablation)DiT-3D baseline · ShapeNet Chair, voxel 32training cost (hours): 91 -> 11Improved
add 3D voxel masking (ablation)DiT-3D baseline · ShapeNet Chair, voxel 321-NNA@CD: 51.99 -> 50.09Improved
add 3D window attention decoder (ablation)FastDiT-3D with voxel masking · ShapeNet Chair, voxel 32training cost (hours): 11 -> 8Improved
add window attention decoder on quality metrics (ablation)FastDiT-3D with voxel masking · ShapeNet Chair, voxel 321-NNA@CD: 50.09 -> 50.35No clear effect
increase random masking ratio to 99% (ablation)FastDiT-3D · ShapeNet Chair, voxel 321-NNA@CD: 51.53 -> 82.35 (95% to 99%)Hurt
use foreground-background aware masking instead of random masking at 95% (ablation)FastDiT-3D · ShapeNet Chair, voxel 321-NNA@CD: 51.53 -> 50.22Improved
increase non-occupied masking ratio from 99% to 100% (ablation)voxel-aware masking · ShapeNet Chair, voxel 321-NNA@CD: 50.09 -> 52.87Hurt
add mixture-of-experts FFN (k=2) in encoder (ablation)FastDiT-3D · ShapeNet Chair, voxel 321-NNA@CD: 52.16 -> 51.72Improved
initialize from ImageNet pre-trained MaskDiT weights (ablation)FastDiT-3D · ShapeNet Chair, voxel 32training cost (hours): 8 -> 7 hours; 1-NNA@CD 50.35 -> 50.39No clear effect
use FastDiT-3D instead of DiT-3D at voxel size 1283D point cloud generation · ShapeNet Chair, voxel 128training cost (A100 hours): 1668 -> 108Improved

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