Research map / Detection and segmentation / Open-set and few-shot detection
Efficient Interactive Annotation of Segmentation Datasets With Polygon-RNN++
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 |
|---|---|---|---|
| replace the VGG encoder with the residual encoder (ablation) | Polygon-RNN · Cityscapes | IoU (%): 61.40 -> 62.16 | Helped |
| add attention (ablation) | Residual Polygon-RNN · Cityscapes | IoU (%): 62.16 -> 65.43 | Helped |
| add reinforcement-learning fine-tuning (ablation) | Residual Polygon-RNN + Attention · Cityscapes | IoU (%): 65.43 -> 67.17 | Helped |
| add the evaluator network (ablation) | Residual Polygon-RNN + RL · Cityscapes | IoU (%): 67.17 -> 70.21 | Helped |
| add GGNN polygon upscaling (ablation) | Polygon-RNN++ · Cityscapes | IoU (%): 70.21 -> 71.38 | Helped |
| add GGNN polygon upscaling (ablation) | Polygon-RNN++ · ADE20K subset | IoU (%): 70.21 -> 71.82 | Helped |
| add GGNN polygon upscaling (ablation) | Polygon-RNN++ · KITTI | IoU (%): 81.40 -> 83.14 | Helped |
| add GGNN polygon upscaling (ablation) | Polygon-RNN++ · Rooftop, Cardiac MR, and ssTEM | IoU (%): 65.03 -> 65.67; 80.55 -> 80.63; 53.77 -> 53.12 | Mixed |
| add 0–5% bounding-box noise | Polygon-RNN++ · Cityscapes | IoU (%): 71.38 -> 70.54 | Hurt |
| add 5–10% bounding-box noise | Polygon-RNN++ · Cityscapes | IoU (%): 71.38 -> 68.07 | Hurt |
| add 10–15% bounding-box noise | Polygon-RNN++ · Cityscapes | IoU (%): 71.38 -> 64.80 | Hurt |
| use Polygon-RNN++ instead of manual annotation | Cityscapes car annotation · Cityscapes car instances | annotation time (s): 39.7 -> 14.7 | Helped |
Closest papers by meaning
- Fast Interactive Object Annotation With Curve-GCN (CVPR 2019 · similarity 0.93)
- Annotating Object Instances With a Polygon-RNN (CVPR 2017 · similarity 0.92)
- Large-Scale Interactive Object Segmentation With Human Annotators (CVPR 2019 · similarity 0.89)
- DatasetGAN: Efficient Labeled Data Factory With Minimal Human Effort (CVPR 2021 · similarity 0.86)
- Training-Free Dataset Pruning for Instance Segmentation (ICLR 2025 · similarity 0.83)
- AGILE3D: Attention Guided Interactive Multi-object 3D Segmentation (ICLR 2024 · similarity 0.83)
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.