Research map / Detection and segmentation / CNN features
Deep Snake for Real-Time 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 |
|---|---|---|---|
| Replace Graph-ResNet with the proposed architecture while retaining graph convolution (ablation) | Curve-gcn + CenterNet contour pipeline · SBD val | APvol: 50.9 -> 52.3 | Helped |
| Add extreme-point octagon initialization before contour deformation (ablation) | Proposed architecture with graph convolution · SBD val | APvol: 52.3 -> 53.6 | Helped |
| Replace graph convolution with circular convolution (ablation) | Deep snake with initial contour proposal · SBD val | APvol: 53.6 -> 54.4 | Helped |
| Use circular rather than graph convolution (ablation) | Contour deformation network · SBD | APvol: Circular conv: 50.6, 54.2, 54.4, 54.0, 53.2; Graph conv: 50.2, 51.5, 53.6, 52.2, 51.6 | Helped |
| Use circular convolution with two iterations instead of graph convolution with three (ablation) | Contour deformation network · SBD | APvol: 54.2 vs 53.6; 0.6 AP | Helped |
| Increase inference iterations beyond three (ablation) | Deep snake contour deformation · SBD | APvol: 54.4 -> 54.0 -> 53.2 | Hurt |
| Train CenterNet with the snake branch rather than alone | CenterNet on KINS · KINS | detection AP: improves by 2.3 AP | Helped |
| Adopt multi-component detection to handle fragmented instances | Deep snake · Cityscapes test | AP: 28.2 | Helped |
Closest papers by meaning
- Object Instance Annotation With Deep Extreme Level Set Evolution (CVPR 2019 · similarity 0.84)
- SharpContour: A Contour-Based Boundary Refinement Approach for Efficient and Accurate Instance Segmentation (CVPR 2022 · similarity 0.83)
- Deep Occlusion-Aware Instance Segmentation With Overlapping BiLayers (CVPR 2021 · similarity 0.83)
- v-CLR: View-Consistent Learning for Open-World Instance Segmentation (CVPR 2025 · similarity 0.81)
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.81)
- Training-Free Dataset Pruning for Instance Segmentation (ICLR 2025 · similarity 0.81)
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.