AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

Semi-Supervised Video Salient Object Detection Based on Uncertainty-Guided Pseudo Labels

Yongri Piao, Chenyang Lu, Miao Zhang, Huchuan Lu

NeurIPS 2022 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Pseudo-labeling for semi-supervised detection · 2 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 CFGMM without adjacent GT reference (CFGMM-O) (ablation)Pseudo label generator · DAVIS train pseudo labelsMAE: 0.016 -> 0.015Improved
add CFGMM with adjacent ground truth reference (ablation)Pseudo label generator · DAVIS train pseudo labelsMAE: 0.016 -> 0.012Improved
add UADDM on top of CFGMM (ablation)Pseudo label generator · DAVIS train pseudo labelsMAE: 0.012 -> 0.007Improved
add CFGMM to baseline (ablation)Saliency model · DAVISMAE: 0.030 -> 0.028Improved
add UADDM on top of CFGMM (ablation)Saliency model · DAVISMAE: 0.028 -> 0.024Improved
add ordinary GAN (ablation)Saliency model · DAVISMAE: 0.024 -> 0.019Improved
replace ordinary GAN with noise-suppressed GAN (ablation)Saliency model · DAVISMAE: 0.019 -> 0.019No clear effect
replace ordinary GAN with noise-suppressed GAN (ablation)Saliency model · MCLMAE: 0.019 -> 0.018Improved
use RGB, flow and fused output streams (ROF) instead of fused only (F) (ablation)Generator · DAVISMAE: 0.043 -> 0.033Improved
increase pseudo labels from 0% to 50% (configuration M) (ablation)Ours · DAVISMAE: 0.034 -> 0.019Improved
increase pseudo labels from 50% to 100% (configuration M) (ablation)Ours · DAVISMAE: 0.019 -> 0.020No clear effect
increase pseudo labels from 50% to 100% (configuration M) (ablation)Ours · FBMSMAE: 0.027 -> 0.031Hurt

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