AI Paper Atlas

Research map / Video / Video segmentation and tracking

Motion-Aware Dynamic Graph Neural Network for Video Compressive Sensing

Ruiying Lu, Ziheng Cheng, Bo Chen, Xin Yuan

TPAMI 2024 · Video › Video segmentation and tracking · approach: Spike and neuromorphic deblurring · 19 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 dynamic walk (DW) sampling (ablation)MadyGraph · grayscale simulationPSNR: 36.57 -> 37.08Improved
add cross-scale (CS) node sampling (ablation)MadyGraph + DW · grayscale simulationPSNR: 37.08 -> 37.21Improved
add motion-aware (MA) information (ablation)MadyGraph + DW + CS · grayscale simulationPSNR: 37.21 -> 37.30Improved
add global knowledge (GK) integration (ablation)MadyGraph + DW + CS + MA · grayscale simulationPSNR: 37.30 -> 37.41Improved
replace fixed-stride sampling with dynamic graph sampling (ablation)MadyGraph · grayscale simulationPSNR: 36.57 -> 37.41Improved
replace self-attention sampling with dynamic graph sampling (ablation)MadyGraph · grayscale simulationPSNR: 37.11 -> 37.41Improved
replace entropy-based sampling with dynamic graph sampling (ablation)MadyGraph · grayscale simulationPSNR: 36.94 -> 37.41Improved
append the graph module to GAP-TVGAP-TV · grayscale simulationPSNR: 26.79 -> 29.50Improved
append the graph module to DeSCIDeSCI · grayscale simulationPSNR: 32.72 -> 34.39Improved
append the graph module to BIRNATBIRNAT · grayscale simulationPSNR: 33.31 -> 33.83Improved
append the graph module to DDUNDDUN · grayscale simulationPSNR: 35.36 -> 36.28Improved
append the graph module to STFormerSTFormer · grayscale simulationPSNR: 36.24 -> 37.28Improved

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