AI Paper Atlas

Research map / Image generation / Image restoration

Correcting Quantization-Induced Gradient Mismatch in Neural Image Compression

Changhao Peng, Yuqi Ye, Wei Gao

AAAI 2026 · Image generation › Image restoration · approach: Learned image compression · 0 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 MSE-guided gradients on mu when |y-mu|<1/2 (ablation)Checkerboard · KodakBD-BR (%): 3.29 -> 1.79Improved
add gradients transfer (k=0.01) (ablation)Checkerboard · KodakBD-BR (%): 3.29 -> 2.32Improved
combine MSE-guided gradients and gradients transfer (ablation)Checkerboard · KodakBD-BR (%): 3.29 -> 1.20Improved
combine both gradient correctionsELIC · KodakBD-BR (%): -8.00 -> -9.55Improved
combine both gradient correctionsMLIC++ · KodakBD-BR (%): -15.26 -> -16.56Improved
combine both gradient correctionsDCAE · KodakBD-BR (%): -17.36 -> -18.16Improved
fine-tune pretrained model without improvement (ablation)Checkerboard · KodakBD-BR (%): 3.29 -> 3.29No clear effect
apply MSE-guided gradients also when round(y-mu)!=0 (ablation)Checkerboard MSE-Grad · KodakBD-BR (%): 1.79 -> 1.84Hurt
lower gradient-transfer k from 0.01 to 0.001 (ablation)Checkerboard GT · KodakBD-BR (%): 2.32 -> 2.85Hurt
raise gradient-transfer k to 0.1 (ablation)Checkerboard GT · KodakBD-BR (%): NA (training NaN)Hurt

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