AI Paper Atlas

Research map / Detection and segmentation / CNN features

ViDSOD-100: A New Dataset and a Baseline Model for RGB-D Video Salient Object Detection

Junhao Lin, Lei Zhu, Jiaxing Shen, Huazhu Fu, Qing Zhang, Liansheng Wang

IJCV 2024 · Detection and segmentation › CNN features · approach: RGB-D salient object detection · 22 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 MEA modules to basic network (ablation)basic · ViDSOD-100MAE: 0.045 -> 0.037Improved
add MDA modules to basic network (ablation)basic · ViDSOD-100MAE: 0.045 -> 0.035Improved
add attention blocks in MDA (ablation)ATF-Net · ViDSOD-100MAE: 0.032 -> 0.028Improved
use ground-truth depth instead of DPT-estimated depth (ablation)ATF-Net · ViDSOD-100MAE: 0.031 -> 0.028Improved
use DPT-B+R+AL depth instead of DPT depth (ablation)ATF-Net · ViDSOD-100MAE: 0.031 -> 0.029Improved
add depth map via MEA+MDA to RGB-only vanilla (ablation)vanilla · ViDSOD-100MAE: 0.066 -> 0.044Improved
add optical flow via MEA+MDA to RGB-only vanilla (ablation)vanilla · ViDSOD-100MAE: 0.066 -> 0.043Improved
add optical flow to depth-equipped network (ablation)vanilla+D-ED · ViDSOD-100MAE: 0.044 -> 0.028Improved
fuse depth by simple concat (ablation)vanilla · ViDSOD-100MAE: 0.066 -> 0.061Improved
fuse depth with UNet-like branch instead of simple concat (ablation)vanilla+D-concat · ViDSOD-100MAE: 0.061 -> 0.054Improved
fuse depth with MEA+MDA instead of UNet-like branch (ablation)vanilla+D-branch · ViDSOD-100MAE: 0.054 -> 0.044Improved
feed depth as extra input to STMSTM · ViDSOD-100MAE: 0.032 -> 0.032No clear effect

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