AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Mask-Guided Spectral-Wise Transformer for Efficient Hyperspectral Image Reconstruction

Yuanhao Cai, Jing Lin, Xiaowan Hu, Haoqian Wang, Xin Yuan, Yulun Zhang, Radu Timofte, Luc Van Gool

CVPR 2022 · Architectures and efficiency › Attention and transformers · approach: Efficient image transformers · 344 citations (OpenAlex)

Check 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 MST-L instead of DGSMPHSI reconstruction · KAIST 10-scene simulation (Avg)PSNR (dB): 32.63 -> 35.18Improved
Add S-MSA to the baseline (ablation)MST-S baseline · KAIST simulationPSNR (dB): 32.29 -> 33.18Improved
Add MM to the S-MSA model (ablation)MST-S baseline with S-MSA · KAIST simulationPSNR (dB): 33.18 -> 34.26Improved
Replace baseline attention with S-MSA (ablation)MST-S baseline · KAIST simulationPSNR (dB): 32.29 -> 33.18Improved
Replace baseline attention with Global MSA (ablation)MST-S baseline · KAIST simulationPSNR (dB): 32.29 -> 32.67Improved
Replace baseline attention with Local W-MSA (ablation)MST-S baseline · KAIST simulationPSNR (dB): 32.29 -> 32.75Improved
Replace baseline attention with Swin W-MSA (ablation)MST-S baseline · KAIST simulationPSNR (dB): 32.29 -> 32.86Improved
Multiply the input by the physical mask (ablation)S-MSA baseline · KAIST simulationPSNR (dB): 33.18 -> 33.57Improved
Use mask-multiplied input together with MM rather than unmodified input with MM (ablation)S-MSA model with MM · KAIST simulationPSNR (dB): 34.26 -> 34.07Hurt

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