AI Paper Atlas

Research map / Detection and segmentation / CNN features

Stacked Cross Refinement Network for Edge-Aware Salient Object Detection

Zhe Wu, Li Su, Qingming Huang

ICCV 2019 · Detection and segmentation › CNN features · approach: RGB-D salient object detection · 486 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 2 CRUs to the two-branch baseline (ablation)two-branch baseline · DUTS-TESTweighted F-measure: 0.752 -> 0.786Improved
stack 4 CRUs instead of 2 (ablation)SCRN · DUTS-TESTweighted F-measure: 0.786 -> 0.803Improved
stack 8 CRUs instead of 4 (ablation)SCRN · DUT-OMRONweighted F-measure: 0.720 -> 0.714Hurt
stack 8 CRUs instead of 4 (ablation)SCRN · DUTS-TESTweighted F-measure: 0.803 -> 0.807No clear effect
make refinement bidirectional rather than segmentation-to-edge only (ablation)SCRN4 · DUTS-TESTweighted F-measure: 0.773 -> 0.803Improved
make refinement bidirectional rather than edge-to-segmentation only (ablation)SCRN4 · DUTS-TESTweighted F-measure: 0.772 -> 0.803Improved
use multiplication for S->E and concatenation for S<-E instead of concatenation both ways (ablation)SCRN4 · DUTS-TESTweighted F-measure: 0.764 -> 0.803Improved
use mixed operations instead of multiplication in both directions (ablation)SCRN4 · DUT-OMRONweighted F-measure: 0.708 -> 0.720Improved
set-to-point style instead of point-to-point (ablation)SCRN4 · DUTS-TESTweighted F-measure: 0.785 -> 0.787Improved
selective set-to-point instead of set-to-point (ablation)SCRN4 · DUTS-TESTweighted F-measure: 0.787 -> 0.803Improved
add residual learning (ablation)SCRN4 SSPS · DUTS-TESTweighted F-measure: 0.802 -> 0.803Improved
add CRUs (4) to baseline, speed cost (ablation)SCRN · FPSFPS: 125 -> 52Hurt

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