AI Paper Atlas

Research map / Detection and segmentation / CNN features

Structured Modeling of Joint Deep Feature and Prediction Refinement for Salient Object Detection

Yingyue Xu, Dan Xu, Xiaopeng Hong, Wanli Ouyang, Rongrong Ji, Min Xu, Guoying Zhao

ICCV 2019 · Detection and segmentation › CNN features · approach: RGB salient object detection · 58 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 prediction-prediction message passing via jointly trained CRF (ablation)Baseline · ECSSDF-measure: 0.884 -> 0.921Improved
add feature-feature message passing (ablation)Baseline · ECSSDF-measure: 0.884 -> 0.904Improved
add feature-prediction messages to feature-feature and prediction-prediction (ablation)CRF model · ECSSDF-measure: 0.921 -> 0.928Improved
replace Dense-CRF post-processing with jointly trained CRF (ablation)Baseline · ECSSDF-measure: 0.902 -> 0.921Improved
joint training of CRF through backpropagation (backbone output) (ablation)Baseline · ECSSDF-measure: 0.884 -> 0.899Improved
emphasize similarity kernel sigma_alpha=10 sigma_beta=10 (ablation)one CRF block at scale 5 · ECSSDF-measure: 0.892 -> 0.910Improved
emphasize proximity kernel (sigma_beta=10, sigma_gamma=10) (ablation)one CRF block at scale 5 · ECSSDMAE: 0.071 -> 0.069Improved
change nu1 and nu2 weights (3,5 / 5,3) (ablation)one CRF block at scale 5 · ECSSDF-measure: 0.892 -> 0.892No clear effect
add CRF at scale 5 to the backbone (ablation)CRF cascade · ECSSDF-measure: 0.884 -> 0.921No clear effect
add vertical flipping augmentation (ablation)CRF block · ECSSDF-measure: 0.910 -> 0.905Hurt
add image noise augmentation (ablation)CRF block · ECSSDF-measure: 0.910 -> 0.896Hurt
add feature-prediction message passing (ablation)CRF cascade · ECSSDtraining lossImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 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