AI Paper Atlas

Research map / Image generation / Super-resolution

MTKD: Multi-Teacher Knowledge Distillation for Image Super-Resolution

Yuxuan Jiang, Chen Feng, Fan Zhang, David Bull

ECCV 2024 · Image generation › Super-resolution · approach: CNN and transformer SR architectures · 42 citations (OpenAlex)

Read the paper

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 three teachers (EDSR, RCAN, SwinIR) instead of SwinIR only (ablation)MTKD with RCAN_lightweight · Urban100 x4PSNR: 26.70 -> 27.08Improved
add EDSR as second teacher alongside SwinIR (ablation)MTKD with RCAN_lightweight · Urban100 x4PSNR: 26.70 -> 26.79Improved
use DCTSwin blocks instead of Mixer Layer blocks in the aggregation network (ablation)MTKD knowledge aggregation · Urban100 x4PSNR: 26.93 -> 27.08Improved
add DCT and IDCT modules to the aggregation network (ablation)MTKD knowledge aggregation · Urban100 x4PSNR: 26.89 -> 27.08Improved
use wavelet-based distillation loss instead of L1 loss (ablation)MTKD · Urban100 x4PSNR: 26.78 -> 27.08Improved
use wavelet loss over DCT-based loss (ablation)MTKD · Urban100 x4PSNR: 26.98 -> 27.08Improved
use wavelet loss over high-frequency-only DWT loss (ablation)MTKD · Urban100 x4PSNR: 26.28 -> 27.08Improved
apply MTKD instead of other KD methods across students and scalesRCAN_lightweight, EDSR_baseline, SwinIR_lightweight · Set5/Set14/BSD100/Urban100 x2 x3 x4PSNR: up to 0.46 dB over second bestImproved
distill with MTKD into RCAN_lightweightRCAN_lightweight · Urban100 x4PSNR: 27.08 vs 26.82 (full RCAN)Improved
gain over other KD methods depends on teacher-student gapMTKD · multiple studentsPSNRMixed

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