AI Paper Atlas

Research map / Detection and segmentation / CNN features

Learning Selective Mutual Attention and Contrast for RGB-D Saliency Detection

Nian Liu, Ni Zhang, Ling Shao, Junwei Han

TPAMI 2021 · Detection and segmentation › CNN features · approach: RGB-D salient object detection · 130 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 mutual attention (MA) (ablation)baseline RGB-D SOD · LFSDmaxF: 0.819 -> 0.864Improved
add mutual attention (MA) (ablation)baseline RGB-D SOD · NJUDS-measure: 0.883 -> 0.898Improved
add mutual attention (MA) (ablation)baseline RGB-D SOD · ReDWeb-SMAE: 0.109 -> 0.115Hurt
add contrast mechanism (MAC) (ablation)MA model · NJUDS-measure: 0.898 -> 0.901Improved
add non-local module (NL) (ablation)MAC model · LFSDS-measure: 0.862 -> 0.874Improved
add contrast mutual decoder (CMD) (ablation)MAC+NL · LFSDmaxF: 0.868 -> 0.881Improved
add selective attention (SMAC+SCMD) (ablation)MAC+NL+CMD · ReDWeb-SMAE: 0.107 -> 0.098Improved
add selective attention (SMAC+SCMD) (ablation)MAC+NL+CMD · LFSDmaxF: 0.881 -> 0.874Hurt
use mutual attention instead of self-attention (ablation)baseline · LFSDmaxF: 0.796 -> 0.864Improved
use MAC+NL instead of S2MA (ablation)SMAC · LFSDS-measure: 0.862 -> 0.874Improved
use high-order SMAC fusion instead of concatenation fusion (ablation)SMAC decoder · ReDWeb-SS-measure: 0.781 -> 0.801Improved
use SMAC instead of summation fusion (ablation)SMAC decoder · NJUDS-measure: 0.899 -> 0.911Improved

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