AI Paper Atlas

Research map / Video / Video segmentation and tracking

Class-agnostic Reconstruction of Dynamic Objects from Videos

Zhongzheng Ren, Xiaoming Zhao, Alex Schwing

NeurIPS 2021 · Video › Video segmentation and tracking · approach: Layered video decomposition and dynamics · 6 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
use REDO's pixel-aligned canonical 4D representation and learned flow instead of OFlowdynamic object reconstruction · SAIL-VOS 3DmIoU: 26.0 -> 31.9Improved
use REDO's pixel-aligned canonical 4D representation and learned flow instead of OFlowdynamic object reconstruction · DeformingThings4D++mIoU: 55.2 -> 57.4Improved
use REDO's pixel-aligned canonical 4D representation and learned flow instead of OFlowdynamic object reconstruction · 3DPWmIoU: 31.5 -> 41.6Improved
replace the transformer temporal aggregator with average pooling (ablation)REDO on SAIL-VOS 3D · SAIL-VOS 3DmIoU: 31.9 -> 28.3Hurt
replace pixel-aligned features with whole-frame feature maps (ablation)REDO on SAIL-VOS 3D · SAIL-VOS 3DmIoU: 31.9 -> 24.1Hurt
remove the temporal correspondence loss (ablation)REDO on SAIL-VOS 3D · SAIL-VOS 3DmACD: 1.47 -> 3.12Hurt

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