AI Paper Atlas

Research map / Image generation / Super-resolution

Feature Representation Matters: End-to-End Learning for Reference-based Image Super-resolution

Yanchun Xie, Jimin Xiao, Mingjie Sun, Chao Yao, Kaizhu Huang

ECCV 2020 · Image generation › Super-resolution · approach: Real-world super-resolution · 44 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 trainable end-to-end SR feature encoder instead of ImageNet VGG features (ablation)E2ENT2 · CUFED5PSNR: 22.85 -> 24.01Improved
use end-to-end trained encoder instead of frozen pre-trained SISR encoder (ablation)E2ENT2 · CUFED5PSNR: 23.46 -> 24.01Improved
use end-to-end trained encoder instead of VGG, feature distance (ablation)E2ENT2 · CUFED5feature distance (lower better): 106.77 -> 25.77Improved
allocate gradient across all three features instead of only the LR input feature (ablation)E2ENT2 · CUFED5PSNR: 23.75 -> 24.01Improved
allocate gradient across all three features instead of only the downsampled reference feature (ablation)E2ENT2 · CUFED5PSNR: 23.67 -> 24.01Improved
allocate gradient across all three features instead of only the HR reference feature (ablation)E2ENT2 · CUFED5PSNR: 23.83 -> 24.01Improved
replace L1 reconstruction loss with MSE loss (ablation)E2ENT2 · CUFED5PSNR: 24.01 -> 24.24Improved
replace L1 reconstruction loss with MSE loss (ablation)E2ENT2 · SUN HaysPSNR: 28.13 -> 28.50Improved
add reference image with end-to-end match and swapOurs-SISR · CUFED5PSNR: 23.75 -> 24.01Improved
add reference image with end-to-end match and swapOurs-SISR · SUN HaysPSNR: 26.72 -> 28.13Improved
use E2ENT2 instead of SRNTT texture transferSRNTT · CUFED5 and SUN HaysPSNR: +0.6 dB on CUFED5, +1.71 dB on SUN HaysImproved
use E2ENT2 instead of SRNTT on stereo dataSRNTT · Flickr1024 testPSNR: 22.02 -> 22.89Improved

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