AI Paper Atlas

Research map / Detection and segmentation / CNN features

JSTASR: Joint Size and Transparency-Aware Snow Removal Algorithm Based on Modified Partial Convolution and Veiling Effect Removal

Wei-Ting Chen, Hao-Yu Fang, Jian-Jiun Ding, Cheng-Che Tsai, Sy-Yen Kuo

ECCV 2020 · Detection and segmentation › CNN features · approach: Context for hard detection conditions · 245 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
add partial shortcut (ablation)JSTASR with partial convolution · SRRS Test A (ablation)PSNR: 24.01 -> 24.86Improved
add multi-convolution (ablation)J+P · SRRS Test A (ablation)PSNR: 24.86 -> 23.02Hurt
add multi-convolution (ablation)J+P · SRRS Test A (ablation)SSIM: 0.87 -> 0.88Improved
add transparency-aware mechanism (ablation)J+P+M · SRRS Test A (ablation)PSNR: 23.02 -> 24.27Improved
add size-aware mechanism without end-to-end veiling removal (ablation)J+P+M+T · SRRS Test A (ablation)PSNR: 24.27 -> 20.32Hurt
add end-to-end veiling effect removal (differentiable DCP) (ablation)All w/o EVER · SRRS Test A (ablation)PSNR: 20.32 -> 25.85Improved
replace DCP with differentiable DCP layer (ablation)dehazing on SOTS indoor · SOTS indoor (dehazing)PSNR: 18.2 -> 24.68Improved
replace PMS-Net with differentiable DCP layer (ablation)dehazing on SOTS indoor · SOTS indoor (dehazing)CIEDE2000: 5.88 -> 4.25Improved
use JSTASR instead of DesnowNetsnow removal · Test B (DesnowNet dataset)PSNR: 25.58 -> 25.69Improved
use JSTASR instead of DesnowNet+MSCNN as preprocessingMask R-CNN detection · Snow-synthesized RTTS object detectionaccuracy (IoU): 0.50 -> 0.72Improved
use JSTASR instead of DesnowNet on SRRS Test Asnow removal · SRRS Test A (w/o MSCNN)PSNR: 16.32 -> 25.62Improved

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