Research map / Detection and segmentation / CNN features
RGB salient object detection
Detect salient objects with attention, edge cues, and recurrent refinement.
61 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in CNN features, keyword cluster “salient · saliency · sod”
State History Benchmarks Tried Baselines Preprints Export
Where it stands
Few recent reports The tracked benchmarks have at most one report from 2025–26 that could beat the old record, too few to say whether they are still moving.
2 papers in 2023–24 and 2 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 0.6% to 0.6%.
Open questions the data points to
- Papers report DCL+ on SOD (MAE) anywhere from 0.126 to 0.153: comparisons against it depend on whose number you take.
No longer reported by papers here: DUT-OMRON (MAE, last in 2020).
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · DHSNet: Deep Hierarchical Saliency Network for Salient Object Detection (CVPR 2016 · 885 citations)
- 2017 · Deeply Supervised Salient Object Detection With Short Connections (CVPR 2017 · 1,206 citations)
- 2018 · PiCANet: Learning Pixel-Wise Contextual Attention for Saliency Detection (CVPR 2018 · 898 citations)
- 2019 · BASNet: Boundary-Aware Salient Object Detection (CVPR 2019 · 1,605 citations)
- 2020 · F³Net: Fusion, Feedback and Focus for Salient Object Detection (AAAI 2020 · 1,082 citations)
- 2021 · Group Collaborative Learning for Co-Salient Object Detection (CVPR 2021 · 106 citations)
- 2022 · Pyramid Grafting Network for One-Stage High Resolution Saliency Detection (CVPR 2022 · 127 citations)
- 2023 · Discriminative Co-Saliency and Background Mining Transformer for Co-Salient Object Detection (CVPR 2023 · 40 citations)
- 2025 · Exploring Salient Object Detection with Adder Neural Networks (AAAI 2025 · 2 citations)
- 2026 · Salient Object Ranking via Cyclical Perception-Viewing Interaction Modeling (ICLR 2026)
Earliest papers here
- DHSNet: Deep Hierarchical Saliency Network for Salient Object Detection (CVPR 2016 · 885 citations)
- Deep Contrast Learning for Salient Object Detection (CVPR 2016 · 739 citations)
- Deep Saliency With Encoded Low Level Distance Map and High Level Features (CVPR 2016 · 515 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
ECSSD · MAE (lower is better)
19 papers, 2016–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 0.06 | Saliency Unified: A Deep Architecture for Simultaneous Eye Fixation Prediction and Salient Object Segmentation (CVPR 2016 · 151 citations) |
| 2017 | 0.054 | Instance-Level Salient Object Segmentation (CVPR 2017 · 303 citations) |
| 2018 | 0.044 | A Bi-Directional Message Passing Model for Salient Object Detection (CVPR 2018 · 512 citations) |
| 2019 | 0.0328 | Pyramid Feature Attention Network for Saliency Detection (CVPR 2019 · 779 citations) |
HKU-IS · MAE (lower is better)
18 papers, 2016–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 0.049 | Deep Contrast Learning for Salient Object Detection (CVPR 2016 · 739 citations) |
| 2017 | 0.039 | Instance-Level Salient Object Segmentation (CVPR 2017 · 303 citations) |
| 2018 | 0.037 | Detect Globally, Refine Locally: A Novel Approach to Saliency Detection (CVPR 2018 · 496 citations) |
| 2019 | 0.03 | An Iterative and Cooperative Top-Down and Bottom-Up Inference Network for Salient Object Detection (CVPR 2019 · 266 citations) |
| 2020 | 0.027 | Label Decoupling Framework for Salient Object Detection (CVPR 2020 · 384 citations) |
PASCAL-S · MAE (lower is better)
17 papers, 2016–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 0.1 | Saliency Unified: A Deep Architecture for Simultaneous Eye Fixation Prediction and Salient Object Segmentation (CVPR 2016 · 151 citations) |
| 2017 | 0.081 | Instance-Level Salient Object Segmentation (CVPR 2017 · 303 citations) |
| 2018 | 0.074 | A Bi-Directional Message Passing Model for Salient Object Detection (CVPR 2018 · 512 citations) |
| 2019 | 0.064 | CapSal: Leveraging Captioning to Boost Semantics for Salient Object Detection (CVPR 2019 · 133 citations) |
| 2020 | 0.06 | Label Decoupling Framework for Salient Object Detection (CVPR 2020 · 384 citations) |
DUT-OMRON · MAE (lower is better)
15 papers, 2016–2020 · not reported since 2020
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 0.07 | Saliency Unified: A Deep Architecture for Simultaneous Eye Fixation Prediction and Salient Object Segmentation (CVPR 2016 · 151 citations) |
| 2017 | 0.069 | Instance-Level Salient Object Segmentation (CVPR 2017 · 303 citations) |
| 2018 | 0.06 | Recurrently Aggregating Deep Features for Salient Object Detection (AAAI 2018 · 169 citations) |
| 2019 | 0.0414 | Pyramid Feature Attention Network for Saliency Detection (CVPR 2019 · 779 citations) |
DUTS-TE · MAE (lower is better)
8 papers, 2019–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 0.04 | An Iterative and Cooperative Top-Down and Bottom-Up Inference Network for Salient Object Detection (CVPR 2019 · 266 citations) |
| 2020 | 0.034 | Label Decoupling Framework for Salient Object Detection (CVPR 2020 · 384 citations) |
HKU-IS · maxF (higher is better)
8 papers, 2016–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 0.904 | Deep Contrast Learning for Salient Object Detection (CVPR 2016 · 739 citations) |
| 2017 | 0.916 | Instance-Level Salient Object Segmentation (CVPR 2017 · 303 citations) |
| 2018 | 0.92 | A Bi-Directional Message Passing Model for Salient Object Detection (CVPR 2018 · 512 citations) |
| 2019 | 0.928 | BASNet: Boundary-Aware Salient Object Detection (CVPR 2019 · 1,605 citations) |
| 2020 | 0.939 | Progressive Feature Polishing Network for Salient Object Detection (AAAI 2020 · 107 citations) |
What has been tried
59 papers here report 601 measured changes (a component added, removed or swapped, compared with a baseline): helped 79%, no clear effect 5%, hurt 14%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use DMM instead of the plain ANN decoder | DUT-O · maxF | 0.647 -> 0.725 | Helped | Exploring Salient Object Detection with Adder Neural Networks (AAAI 2025 · 2 citations) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
Recent preprints
Loading recent arXiv preprints…
Use it
Download BibTeX of 22 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- RGB-D salient object detection (47 papers)
Elsewhere in the topic
- Context modeling for semantic segmentation (241 papers)
- Region-based and query-based detectors (183 papers)
- Local features and context in CNNs (170 papers)
- Context for hard detection conditions (133 papers)
- Domain-specific segmentation and detection (118 papers)
- Convolution-transformer backbones (101 papers)
- Box regression and loss design (94 papers)
- Efficient DETR query design (85 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).