AI Paper Atlas

Research map / Image generation / Image restoration

Practical Deep Raw Image Denoising on Mobile Devices

Yuzhi Wang, Haibin Huang, Qin Xu, Jiaming Liu, Yiqun Liu, Jue Wang

ECCV 2020 · Image generation › Image restoration · approach: Raw and low-light noise modeling · 133 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
train on SID-based synthetic data matching sensor noise instead of SIDDdenoising network · Real phone test setPSNRImproved
use mobile-friendly network instead of UNet-36Gdenoiser · Snapdragon 855 GPUms per megapixel: 383.9 -> 70.70Improved
use mobile-friendly network instead of UNet-21Gdenoiser · Snapdragon 855 GPUms per megapixel: 292.0 -> 70.70Improved
use accurate noise parameters instead of scaling by 0.5 (ablation)synthetic noise training · Real phone test setPSNR (dB): 36.87384014 -> 39.96560621Improved
use accurate noise parameters instead of scaling by 2.0 (ablation)synthetic noise training · Real phone test setPSNR (dB): 39.68042145 -> 39.96560621Improved
train on SID synthesized data instead of test-set overfitting (ablation)denoising network · Real phone test setPSNR (dB): 40.06 -> 39.76Mixed
add k-Sigma Transform instead of iso-augmentation (ablation)denoising network · Multi-ISO test setPSNR (dB): 39.56662978 -> 39.76Improved
add k-Sigma Transform instead of concat-variance (ablation)denoising network · Multi-ISO test setPSNR (dB): 39.653042266876 -> 39.76Improved
add explicit noise variance channels instead of blind iso-augmentation (ablation)denoising network · Multi-ISO test setPSNR (dB): about 0.09 dB higherImproved
train one multi-ISO model with k-Sigma Transform instead of single-ISO 1600 model (ablation)denoising network · Multi-ISO test setPSNR (dB): 35.73569171524029 -> 39.76Improved
train one model with k-Sigma Transform instead of single-ISO model matched to test ISO (ablation)denoising network · ISO-1600 test imagesPSNR (dB): 41.48 -> 41.48No clear effect
use k-Sigma Transform instead of single-ISO 1600 model at high ISO (ablation)denoising network · ISO-6400 test imagesPSNR (dB): 28.86 -> 36.94Improved

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