Research map / Detection and segmentation / CNN features
Pixelwise Instance Segmentation With a Dynamically Instantiated Network
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 the Global unary term to the Box term (ablation) | Instance segmentation network, end-to-end trained · VOC 2012 Validation Set | AP_r at IoU 0.9: 24.6 -> 25.5 | Helped |
| Add the Shape unary term to Box+Global (ablation) | Instance segmentation network, end-to-end trained · VOC 2012 Validation Set | AP_r vol: Training with the “Global” and “Shape” terms shows larger improvements of 2.1% and 2.3% respectively. | Helped |
| Train the full network end-to-end instead of piecewise (ablation) | Instance segmentation network with Box, Global, and Shape terms · VOC 2012 Validation Set | AP_r vol: 2.3% | Helped |
| Add the Global unary term to the Box term (ablation) | Piecewise-trained instance segmentation network · VOC 2012 Validation Set | AP_r vol: 54.9 -> 54.6 | Hurt |
| Add the Shape unary term to Box+Global (ablation) | Piecewise-trained instance segmentation network · VOC 2012 Validation Set | AP_r vol: The “Shape” term’s improvement in the AP_r vol is primarily due to an improvement in the AP_r at low thresholds. | Helped |
| Finetune the semantic segmentation network for instance segmentation (ablation) | Semantic segmentation network · VOC Validation Set | Mean IoU: 74.2 -> 75.1 | Helped |
| Finetune the semantic segmentation network for instance segmentation (ablation) | Semantic segmentation network · SBD Validation Set | Mean IoU: 71.5 -> 72.5 | Helped |
| Use the proposed globally coherent instance segmentation system | Instance segmentation on VOC · VOC 2012 Validation Set | AP_r at IoU 0.9: 6.6% | Helped |
| Use the proposed end-to-end instance segmentation system | Instance segmentation on SBD · SBD Validation Set | AP_r at IoU 0.7: 1.5% | Helped |
| Use the proposed instance segmentation system instead of MNC | Instance segmentation on SBD · SBD Validation Set | Matching IoU: 8.3% | Helped |
| Use instance segmentation finetuning with the proposed network | Instance segmentation on Cityscapes · Cityscapes Test Set | AP: 20.0; 17.4 | Helped |
Closest papers by meaning
- Instance-Aware Semantic Segmentation via Multi-Task Network Cascades (CVPR 2016 · similarity 0.92)
- InstanceCut: From Edges to Instances With MultiCut (CVPR 2017 · similarity 0.91)
- Boundary-Aware Instance Segmentation (CVPR 2017 · similarity 0.90)
- P2Seg: Pointly-supervised Segmentation via Mutual Distillation (ICLR 2024 · similarity 0.89)
- Semantic-aware SAM for Point-Prompted Instance Segmentation (CVPR 2024 · similarity 0.89)
- Training-Free Dataset Pruning for Instance Segmentation (ICLR 2025 · similarity 0.88)
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.