AI Paper Atlas

Research map / Point clouds / Point cloud classification

DynaTok: Token-Based 4D Reconstruction from Partial Point Clouds

Weirong Chen, Keisuke Tateno, Hidenobu Matsuki, Michael Niemeyer, Daniel Cremers, Federico Tombari

ICML 2026 · Point clouds › Point cloud classification · approach: Scene flow and dynamic point clouds · 0 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 cross-frame global attention (joint spatiotemporal alignment) to the token encoder (ablation)DynaTok · DT4D-AAccuracy (Acc.): 0.276 -> 0.027Improved
add cross-frame global attention (ablation)DynaTok · DT4D-ANormal Consistency: 0.531 -> 0.877Improved
increase tokens per frame from 128 to 256 (ablation)DynaTok · DT4D-ACompleteness: 0.032 -> 0.026Improved
increase tokens per frame from 256 to 512 (ablation)DynaTok · DT4D-AAccuracy / Completeness / NC: Acc 0.027 -> 0.026; Comp 0.026 -> 0.030; NC 0.877 -> 0.881Mixed
increase number of training views from 2 to 8 (ablation)DynaTok · DT4D-ACompleteness: 0.035 -> 0.026Improved
use DynaTok temporal aggregation instead of frame-wise 3D latent baselinesKubric foreground · Kubric, Foreground + BackgroundCompleteness: 0.020 -> 0.014 (vs TripoSG)Improved
use DynaTok instead of Shape2VecSet on accuracyKubric Foreground + Background · Kubric, Foreground + BackgroundAccuracy: 0.009 -> 0.010No clear effect
use DynaTok instead of Motion2VecSetsDT4D-A Unseen Motion · DT4D-A, Unseen MotionCompleteness: 0.060 -> 0.021Improved
use DynaTok instead of Motion2VecSets for efficiency16-frame scene inference · Model complexityParams / time per scene: 423M -> 19M params; ~45s -> ~22sImproved

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