Research map / Detection and segmentation / CNN features
DoNet: Deep De-Overlapping Network for Cytology Instance Segmentation
Hao Jiang, Rushan Zhang, Yanning Zhou, Yumeng Wang, Hao Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Use DoNet's decompose-and-recombine network instead of competing instance-segmentation methods | cytology cell instance segmentation · ISBI2014 | mAP: 2.68% | Improved |
| Use DoNet's decompose-and-recombine network instead of competing instance-segmentation methods | cytology cell instance segmentation · CPS | mAP: 1.85% | Improved |
| Add DRM to the Mask R-CNN baseline (ablation) | Mask R-CNN · ISBI2014 | mAP: 3.25% | Improved |
| Add DRM to the Mask R-CNN baseline (ablation) | Mask R-CNN · CPS | mAP | Hurt |
| Add CRM after DRM to recombine sub-regions (ablation) | DoNet · ISBI2014 | mAP: 7.34% | Improved |
| Add CRM after DRM to recombine sub-regions (ablation) | DoNet · CPS | mAP: 1.74% | Improved |
| Add MRP to reweight proposal features using cytoplasm predictions (ablation) | DoNet · ISBI2014 | mAP: 0.59% | Improved |
| Add MRP to reweight proposal features using cytoplasm predictions (ablation) | DoNet · CPS | mAP: 0.31% | Improved |
| Add intersection head Ho (ablation) | DRM with instance head Hi · ISBI2014 | mAP (average): 2.20% | Improved |
| Add complement head Hm (ablation) | DRM with Hi and Ho · ISBI2014 | mAP (average): 2.64% | Improved |
| Add Fusion Unit to integrate overlapping and non-overlapping region features (ablation) | CRM · ISBI2014 | mAP (average): 1.52% | Improved |
| Add consistency regularization between recombined masks and merged sub-regions (ablation) | CRM · ISBI2014 | mAP (cytoplasm): 0.91% | Improved |
Closest papers by meaning
- Disco: Densely-overlapping Cell Instance Segmentation via Adjacency-aware Collaborative Coloring (ICLR 2026 · similarity 0.86)
- COIN: Confidence Score-Guided Distillation for Annotation-Free Cell Segmentation (ICCV 2025 · similarity 0.85)
- The Four Color Theorem for Cell Instance Segmentation (ICML 2025 · similarity 0.83)
- Weakly Supervised Gland Segmentation with Class Semantic Consistency and Purified Labels Filtration (AAAI 2025 · similarity 0.82)
- Beyond ViT Tokens: Masked-Diffusion Pretrained Convolutional Pathology Foundation Model for Cell-Level Dense Prediction (NeurIPS 2026 · similarity 0.81)
- Mudslide: A Universal Nuclear Instance Segmentation Method (CVPR 2024 · 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.
