AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Spectrum-Aware and Transferable Architecture Search for Hyperspectral Image Restoration

Wei He, Quanming Yao, Naoto Yokoya, Tatsumi Uezato, Hongyan Zhang, Liangpei Zhang

ECCV 2022 · Architectures and efficiency › Neural architecture search · approach: Bi-level differentiable NAS · 17 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 STAS search space instead of DARTS space (ablation)STAS · CAVE denoising (noise 50)PSNR: 32.18 -> 36.80Improved
use STAS search space instead of HSI-DeNet baseline (ablation)STAS · CAVE denoising (noise 50)PSNR: 33.85 -> 36.80Improved
add global skip connection and local skips (ablation)STAS · CAVE denoising (noise 50)PSNR: 35.22 -> 36.80Improved
use learned adaptive operations instead of random selection (ablation)STAS · CAVE denoising (noise 50)PSNR: 35.25 -> 36.80Improved
use learned operations instead of spectral-only convolution (ablation)STAS · CAVE denoising (noise 50)PSNR: 32.81 -> 36.80Improved
use learned operations instead of spatial-only convolution (ablation)STAS · CAVE denoising (noise 50)PSNR: 34.72 -> 36.80Improved
optimize STAS with DARTS- instead of default (ablation)STAS · CAVE denoising (noise 50)PSNR: 36.80 -> 36.42Hurt
optimize STAS with SPOS instead of default (ablation)STAS · CAVE denoising (noise 50)PSNR: 36.80 -> 36.89No clear effect
transfer CAVE-searched STASC to Pavia instead of searching on Pavia (ablation)STAS · Pavia denoising (noise 50)PSNR: 34.96 -> 33.10Hurt
search per noise level instead of one mixed-noise search (ablation)STAS · CAVE denoisingPSNR: 36.80 -> 36.67No clear effect
use STASC instead of TSA for KAIST reconstructionSTAS · KAIST averagePSNR: 30.15 -> 31.14Improved

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