Research map / Image generation / Fast diffusion sampling
Style-Extracting Diffusion Models for Semi-Supervised Histopathology Segmentation
Mathias Öttl, Frauke Wilm, Jana Steenpass, Jingna Qiu, Matthias Rübner, Prof Arndt Hartmann, Matthias W. Beckmann, Peter Fasching, Andreas K Maier, Ramona Erber, Bernhard Kainz, Katharina Breininger
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 |
|---|---|---|---|
| add synthetic images from STEDM with multi-patch style conditioning to segmentation training | U-Net MiT-B2 segmentation · HER2 (4 WSI) | Mean IoU: 74.17 -> 78.81 | Improved |
| add STEDM synthetic images | U-Net MiT-B2 segmentation · HER2 (4 WSI) | IoU variance between patients: 1.84 -> 0.24 | Improved |
| add STEDM multi-patch synthetic images | U-Net MiT-B2 segmentation · CATCH (7 WSI) | Mean IoU: 85.71 -> 85.09 (multi-patch); 85.71 -> 86.61 (nearby) | Mixed |
| add STEDM synthetic images to flower segmentation | U-Net MiT-B2 segmentation · Flowers (960 images) | Mean IoU: 87.80 -> 86.49 | Hurt |
| add synthetic images from neural style transfer | U-Net MiT-B2 segmentation · HER2 (24 WSI) | Mean IoU: 78.88 -> 78.67 | No clear effect |
| add semantic-conditioned diffusion model images without style conditioning | U-Net MiT-B2 segmentation · HER2 (12 WSI) | Mean IoU: 74.63 -> 77.28 | Improved |
| use sViT style encoder instead of Swin style encoder with aggregation block (ablation) | STEDM · HER2 (4 WSI) | Mean IoU: 78.81 -> 76.55 | Hurt |
| style conditioning on image generation FID | STEDM · Flowers (480 train images) | FID: 80.78 -> 56.78 | Improved |
| style conditioning on Inception Score | STEDM · Flowers | IS: 4.36 -> 3.50 (960 images) | No clear effect |
| use multi-patch style sampling with aggregation block instead of single nearby patch (ablation) | STEDM · HER2 (4 WSI) | Mean IoU: 77.61 -> 78.81 | Improved |
Closest papers by meaning
- Semantic and Visual Crop-Guided Diffusion Models for Heterogeneous Tissue Synthesis in Histopathology (NeurIPS 2025 · similarity 0.90)
- STREAM: Stochastic Riemannian Flow Matching with Anisotropic Decoder for Digital Histopathology Image Generation (NeurIPS 2026 · similarity 0.88)
- L-Diffusion: Laplace Diffusion for Efficient Pathology Image Segmentation (ICML 2025 · similarity 0.88)
- Dynamic Entity-Masked Graph Diffusion Model for Histopathology Image Representation Learning (AAAI 2025 · similarity 0.88)
- Prototype-Based Image Prompting for Weakly Supervised Histopathological Image Segmentation (CVPR 2025 · similarity 0.87)
- Beyond ViT Tokens: Masked-Diffusion Pretrained Convolutional Pathology Foundation Model for Cell-Level Dense Prediction (NeurIPS 2026 · similarity 0.87)
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.
