Research map / Detection and segmentation / CNN features
S4Net: Single Stage Salient-Instance Segmentation
Check 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 |
|---|---|---|---|
| Add ternary RoIMasking with sign-flipped surrounding context (ablation) | S4Net · Salient instance segmentation validation set | mAP^0.7: 61.5% -> 63.6% | Helped |
| Add ternary rather than binary RoIMasking (ablation) | S4Net · Salient instance segmentation validation set | mAP^0.7_O: 49.4% -> 51.5% | Helped |
| Use expanded binary context without sign flipping (ablation) | S4Net RoIMasking · Salient instance segmentation | mAP | Hurt |
| Set context expansion coefficient α to 1/3 (ablation) | S4Net ternary RoIMasking · Salient instance segmentation validation set | mAP^0.5: 85.9% -> 86.7% | Helped |
| Set context expansion coefficient α to 1/3 rather than 1 (ablation) | S4Net ternary RoIMasking · Salient instance segmentation validation set | mAP^0.7: 62.0% -> 63.6% | Helped |
| Increase proposal count beyond 20 (ablation) | S4Net · Salient instance segmentation | performance: around 1.5% | No clear effect |
| Replace ResNet-50 with ResNet-101 (ablation) | S4Net · Salient instance segmentation benchmark | mAP@0.7: 63.6% -> 66.8% | Helped |
| Replace ResNet-50 with ResNet-101 (ablation) | S4Net · Salient instance segmentation benchmark | Speed (FPS): 40.0 -> 33.3 | Hurt |
| Replace ResNet-50 with MobileNet (ablation) | S4Net · Salient instance segmentation benchmark | Speed (FPS): 40.0 -> 90.9 | Helped |
| Use salient-instance cues instead of saliency-map cues | DeepLab-VGG16 · PASCAL VOC val | semantic segmentation performance: 52.6% -> 57.4% | Helped |
| Replace MSRNet with S4Net | Salient instance segmentation · Salient instance segmentation test set | mAP^0.5: 65.3% -> 86.7% | Helped |
| Replace MSRNet with S4Net | Salient instance segmentation · Salient instance segmentation test set | mAP^0.7: 52.3% -> 63.6% | Helped |
Closest papers by meaning
- Instance-Level Salient Object Segmentation (CVPR 2017 · similarity 0.89)
- Semantic-aware SAM for Point-Prompted Instance Segmentation (CVPR 2024 · similarity 0.87)
- The Devil is in the Object Boundary: Towards Annotation-free Instance Segmentation using Foundation Models (ICLR 2024 · similarity 0.86)
- P2Seg: Pointly-supervised Segmentation via Mutual Distillation (ICLR 2024 · similarity 0.85)
- SAI3D: Segment Any Instance in 3D Scenes (CVPR 2024 · similarity 0.85)
- Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 2024 · similarity 0.85)
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.