AI Paper Atlas

Research map / Detection and segmentation / CNN features

A Stagewise Refinement Model for Detecting Salient Objects in Images

Tiantian Wang, Ali Borji, Lihe Zhang, Pingping Zhang, Huchuan Lu

ICCV 2017 · Detection and segmentation › CNN features · approach: RGB salient object detection · 459 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 a second refinement stage (ablation)1-stage baseline · ECSSDF-measure: 0.843 -> 0.867Improved
add a second refinement stage (ablation)1-stage baseline · DUTSF-measure: 0.682 -> 0.706Improved
add pyramid pooling module to 2-stage (ablation)2-stage · DUTSF-measure: 0.706 -> 0.734Improved
add pyramid pooling module to 2-stage (ablation)2-stage · ECSSDF-measure: 0.867 -> 0.878Improved
add pyramid pooling module to 3-stage (ablation)3-stage · ECSSDF-measure: 0.879 -> 0.892Improved
add pyramid pooling module to 3-stage (ablation)3-stage · DUTSF-measure: 0.735 -> 0.757Improved
go from 3 to 4 stages with ppm (ablation)3-stage+ppm · ECSSDF-measure: 0.892 -> 0.882Hurt
go from 3 to 4 stages with ppm (ablation)3-stage+ppm · DUTSF-measure: 0.757 -> 0.742No clear effect
use stage-wise refinement instead of ResNet-50 FCN with skip connections (ablation)ResNet-50 FCN · ECSSDF-measure: 0.864 -> 0.892Improved
use stage-wise refinement instead of ResNet-50 FCN with skip connections (ablation)ResNet-50 FCN · DUTSF-measure: 0.689 -> 0.757Improved
add auxiliary losses on intermediate stages (ablation)3-stage network · DUT-OMRON and DUTSF-measureImproved
train each stage independently first then jointly instead of direct end-to-end training (ablation)3-stage network · DUT-OMRON and DUTSF-measureHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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