AI Paper Atlas

Research map / Detection and segmentation / CNN features

Specificity-Preserving RGB-D Saliency Detection

Tao Zhou, Huazhu Fu, Geng Chen, Yi Zhou, Deng-Ping Fan, Ling Shao

ICCV 2021 · Detection and segmentation › CNN features · approach: RGB-D salient object detection · 303 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 CIM instead of simple concatenation fusion (ablation)SP-Net · NJU2KS-measure: .916 -> .925Improved
add cross-modal feature enhancement part of CIM on top of adaptive fusion (ablation)SP-Net · NJU2KS-measure: .919 -> .925Improved
add adaptive feature fusion part of CIM on top of cross-modal enhancement (ablation)SP-Net · NJU2KS-measure: .921 -> .925Improved
add cross-level propagation of CIM output to the next layer (ablation)SP-Net · NJU2KS-measure: .924 -> .925Improved
add the MFA module integrating modality-specific decoder features (ablation)SP-Net · NJU2KS-measure: .918 -> .925Improved
use MFA instead of cross-modal feature enhancement fusion (ablation)SP-Net · NJU2KS-measure: .924 -> .925Improved
use MFA instead of simple concatenation fusion (ablation)SP-Net · NJU2KS-measure: .921 -> .925Improved
add the two modality-specific decoders (ablation)SP-Net · NJU2KS-measure: .913 -> .925Improved
use shared decoder output instead of combining the two modality-specific decoders (ablation)SP-Net · NJU2KS-measure: .916 -> .925Improved
use five CIMs instead of one CIM (last layer only) (ablation)SP-Net · NJU2KS-measure: .918 -> .925Improved
use five CIMs instead of three CIMs (ablation)SP-Net · NJU2KS-measure: .920 -> .925Improved
use Res2Net-50 backbone instead of ResNet-50 (ablation)SP-Net · NJU2KS-measure: .922 -> .925Improved

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