Research map / Detection and segmentation / Open-set and few-shot detection
Stochastic Segmentation Networks: Modelling Spatially Correlated Aleatoric Uncertainty
Miguel Monteiro, Loic Le Folgoc, Daniel Coelho de Castro, Nick Pawlowski, Bernardo Marques, Konstantinos Kamnitsas, Mark van der Wilk, Ben Glocker
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 |
|---|---|---|---|
| model correlated logit uncertainty with a low-rank covariance instead of independent pixel-wise uncertainty (ablation) | SSN on LIDC-IDRI · LIDC-IDRI, one annotation set | D²_GED: 0.734 ± 0.009 -> 0.365 ± 0.005 | Improved |
| model correlated logit uncertainty with a low-rank covariance instead of independent pixel-wise uncertainty (ablation) | SSN on LIDC-IDRI · LIDC-IDRI, four annotation sets | DSC (%): 37.0 ± 0.4 -> 43.6 ± 0.4 | Improved |
| train with all four expert annotations rather than one annotation set (ablation) | SSN low-rank · LIDC-IDRI | DSC (%): 40.7 ± 0.4 -> 43.6 ± 0.4 | Improved |
| increase label-map patch size from 30 mm³ to 60 mm³ (ablation) | SSN low-rank · BraTS 2017 | DSC_WT (%): 88.0 ± 1.3 -> 88.7 ± 1.3 | No clear effect |
| increase label-map patch size from 30 mm³ to 60 mm³ (ablation) | SSN low-rank · BraTS 2017 | D²_GED: 0.635 ± 0.029 -> 0.689 ± 0.031 | No clear effect |
| vary covariance rank above rank one (ablation) | SSN low-rank · BraTS 2017 | performance metrics | No clear effect |
| use low-rank covariance rather than diagonal covariance (ablation) | SSN on the toy problem · Toy problem | log-likelihood: -4.87 -> -0.93 | Improved |
| use stochastic low-rank logits rather than a deterministic network | SSN on BraTS 2017 · BraTS 2017 | DSC_WT (%): 88.2 ± 1.3 -> 88.0 ± 1.3 | No clear effect |
Closest papers by meaning
- Structuring Uncertainty for Fine-Grained Sampling in Stochastic Segmentation Networks (NeurIPS 2022 · similarity 0.91)
- Flow Stochastic Segmentation Networks (ICCV 2025 · similarity 0.89)
- Modeling Multimodal Aleatoric Uncertainty in Segmentation with Mixture of Stochastic Experts (ICLR 2023 · similarity 0.89)
- Tyche: Stochastic In-Context Learning for Medical Image Segmentation (CVPR 2024 · similarity 0.86)
- DyCON: Dynamic Uncertainty-aware Consistency and Contrastive Learning for Semi-supervised Medical Image Segmentation (CVPR 2025 · similarity 0.86)
- Multi-view Evidential Learning-based Medical Image Segmentation (AAAI 2025 · similarity 0.86)
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.
