AI Paper Atlas

Research map / Video / Video segmentation and tracking

Progressive Fusion Video Super-Resolution Network via Exploiting Non-Local Spatio-Temporal Correlations

Peng Yi, Zhongyuan Wang, Kui Jiang, Junjun Jiang, Jiayi Ma

ICCV 2019 · Video › Video segmentation and tracking · approach: Space-time video super-resolution · 316 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 one non-local residual block (r=1) as preprocessing (ablation)VSR · training Vid validationPSNR (dB): about 0.14 dBImproved
add SPMC motion compensation as preprocessing (ablation)VSR · training validationPSNR (dB): about 0.04 dBNo clear effect
add STMC motion compensation as preprocessing (ablation)VSR · training validationPSNR (dB)No clear effect
increase non-local down-scale from r=1 to r=2 (ablation)VSR-NL · training validationPSNR (dB)No clear effect
share parameters across channels (CCPS) (ablation)PFS · fusion comparisonPSNR (dB): 31.21 -> 31.17Hurt
share parameters across blocks (CBPS) (ablation)DFS · fusion comparisonPSNR (dB): 30.98 -> 30.75Hurt
share parameters across blocks (CBPS) (ablation)SFS · fusion comparisonPSNR (dB): 30.86 -> 30.73Hurt
share parameters across blocks (CBPS) (ablation)3DFS · fusion comparisonPSNR (dB): 31.18 -> 30.82Hurt
use progressive fusion PFS instead of 3D convolution fusion (ablation)3DFS · fusion comparisonPSNR (dB): 31.18 -> 31.21Improved
use direct fusion instead of slow fusion (ablation)SFS · fusion comparisonPSNR (dB): 30.86 -> 30.98Improved
increase input frames from 3 to 5 to 7 (ablation)PFS-PS · fusion comparisonPSNR (dB)Improved
use PFNL instead of DUF 52LDUF 52L · 10 sequences 1272x720PSNR (dB): about 0.96 dBImproved

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