Research map / Detection and segmentation / Weakly supervised detection and segmentation
Pixel Is All You Need: Adversarial Trajectory-Ensemble Active Learning for Salient Object Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use ATAL-selected point labels, with 10 annotated pixels per image | MINet, F3Net, and PFSN · DUTS-O, DUTS-TE, ECSSD, HKU-IS, and PASCAL-S | maxF | Mixed |
| Add AATU to DEN (ablation) | F3Net active learning · DUT-OMRON | maxFβ: .7562 -> .7716 | Helped |
| Add AATU to DEN (ablation) | F3Net active learning · DUTS-TE | maxFβ: .8261 -> .8383 | Helped |
| Replace DEN with TEUE (ablation) | F3Net active learning · DUT-OMRON | maxFβ: .7716 -> .7708 | No clear effect |
| Replace DEN with TEUE (ablation) | F3Net active learning · DUTS-TE | maxFβ: .8383 -> .8374 | No clear effect |
| Add cyclic cosine learning strategy to trajectory ensemble (ablation) | TUTE · DUT-OMRON | maxFβ: .7336 -> .7708 | Helped |
| Add cyclic cosine learning strategy to trajectory ensemble (ablation) | TUTE · DUTS-TE | maxFβ: .7964 -> .8374 | Helped |
| Add RDS instead of top-k sampling (ablation) | AATU+TEUE active learning · DUT-OMRON | maxFβ: .7708 -> .7814 | Helped |
| Add RDS instead of top-k sampling (ablation) | AATU+TEUE active learning · DUTS-TE | maxFβ: .8374 -> .8479 | Helped |
| Use superpixel blocks as the labeled unit instead of individual pixels (ablation) | AATU+TEUE+RDS · DUT-OMRON | maxFβ: .7814 -> .7908 | Helped |
| Use superpixel blocks as the labeled unit instead of individual pixels (ablation) | AATU+TEUE+RDS · DUTS-TE | maxFβ: .8479 -> .8624 | Helped |
| Use trajectory snapshots without CCLS rather than DEN (ablation) | Trajectory ensemble uncertainty estimation · DUT-OMRON | maxFβ: .7716 -> .7336 | Hurt |
Closest papers by meaning
- Few-Cost Salient Object Detection with Adversarial-Paced Learning (NeurIPS 2020 · similarity 0.86)
- Weakly Supervised Salient Object Detection Using Image Labels (AAAI 2018 · similarity 0.85)
- Weakly-Supervised Salient Object Detection Using Point Supervision (AAAI 2022 · similarity 0.85)
- Saliuitl: Ensemble Salience Guided Recovery of Adversarial Patches against CNNs (CVPR 2025 · similarity 0.85)
- AS-Det: Active Sampling for Adaptive 3D Object Detection in Point Clouds (AAAI 2025 · similarity 0.83)
- Plug and Play Active Learning for Object Detection (CVPR 2024 · similarity 0.82)
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.