AI Paper Atlas

Research map / Point clouds / Point cloud classification

OctFormer: Efficient Octree-Based Transformer for Point Cloud Compression with Local Enhancement

Mingyue Cui, Junhua Long, Mingjian Feng, Boyang Li, Huang Kai

AAAI 2023 · Point clouds › Point cloud classification · approach: Scene flow and dynamic point clouds · 40 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
Increase the non-overlapped context window size from N=32 to N=1024 (ablation)OctFormer · ScanNet, D=9Bpp: 6.416 -> 6.021Improved
Add OctLeFF in place of the vanilla feed-forward network (ablation)OctFormer · ScanNet, D=9Bpp: 6.212 -> 6.144Improved
Add OctPEG instead of absolute positional encoding (ablation)OctFormer · ScanNet, D=9Bpp: 6.212 -> 6.161Improved
Use absolute positional encoding (ablation)OctFormer · ScanNet, D=9Bpp: 6.212 -> 6.213No clear effect
Add both OctLeFF and OctPEG (ablation)OctFormer · ScanNet, D=9Bpp: 6.212 -> 6.135Improved
Use OctLeFF and OctPEG rather than APE alone (ablation)OctFormer · ScanNet, D=9parameters: 5.66 M -> 4.28 MImproved
Use OctFormer instead of OctAttention without sibling-occupancy featuresOctAttention · SemanticKITTI, 1,000 dummy octree nodesdecode time (s): 2.060 -> 0.007Improved
Use OctFormer instead of VoxelContext-NetVoxelContext-Net · SemanticKITTI, D=12Bpp: 4.497 -> 3.708Improved
Use OctFormer instead of VoxelContext-NetVoxelContext-Net · SemanticKITTI, D=10Bpp: 0.951 -> 0.963Hurt
Use OctFormer instead of OctAttentionOctAttention · SemanticKITTI, D=12Bpp: 3.740 -> 3.708Improved

Later papers that compared with this method

6 later accepted papers report OctFormer as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 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