AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

GridFormer: Residual Dense Transformer with Grid Structure for Image Restoration in Adverse Weather Conditions

Tao Wang, Kaihao Zhang, Ziqian Shao, Wenhan Luo, Björn Stenger, Tong Lü, Tae‐Kyun Kim, Wei Liu, Hongdong Li

IJCV 2024 · Architectures and efficiency › Attention and transformers · approach: Efficient image transformers · 253 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 feature sampling and channel split in CESA (ablation)GridFormer · RainDropPSNR: 0.89 dB and 1.16 dB improvementsImproved
add feature sampling and channel split in CESA (ablation)GridFormer · SOTS-IndoorPSNR: 0.92 dB and 2.70 dB improvementsImproved
add local enhancement in CESA (ablation)GridFormer · RainDropPSNR: 31.98 -> 32.57Improved
add local enhancement in CESA (ablation)GridFormer · SOTS-IndoorPSNR: 40.51 -> 41.84Improved
exchange Values instead of Queries between streams (ablation)GridFormer · RainDropPSNR: 31.81 -> 32.57Improved
pooling strides r=[4,2,2] instead of [2,2,2] (ablation)GridFormer · SOTS-IndoorPSNR: 33.81 -> 40.51Improved
pooling strides r=[4,2,2] instead of [4,4,4] (ablation)GridFormer · RainDropPSNR: 31.62 -> 31.98Improved
add dense connections in RDTB (ablation)GridFormer · RainDropPSNR: 31.48 -> 32.07Improved
add local skip connection in RDTB (ablation)GridFormer · SOTS-IndoorPSNR: 40.87 -> 41.84Improved
add local fusion with 1x1 convolution in RDTB (ablation)GridFormer · RainDropPSNR: 32.07 -> 32.05No clear effect
increase grid columns from c=3 to c=6 at r=3 (ablation)GridFormer · RainDropPSNR: 31.67 -> 32.57Improved
increase grid rows from r=3 to r=4 (ablation)GridFormer · RainDropPSNR: 32.57 -> 32.05Hurt

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