Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Apr 9, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

OOD-SEG: Exploiting out-of-distribution detection techniques for learning image segmentation from sparse multi-class

Junwen Wang¹, Zhonghao Wang¹, Oscar MacCormac²

¹School of Biomedical Engineering & Imaging Sciences, King's College London, UK.

Medical Image Analysis

|April 7, 2026

Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Label tree semantic losses for rich multi-class medical image segmentation.

Frontiers in artificial intelligence·2026

Same author

Lower cranial nerve responses after laryngeal stimulation in posterior fossa surgery: a study of medullary complex motor functions.

Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology·2026

Same author

Complement C5a/C5aR1 pathway facilitates glioblastoma progression via fostering glioma stem cell-macrophage symbiosis.

Journal of neuroinflammation·2026

Same author

Evaluation of Six Cutting-edge Large Language Models in Answering a Set of Open-ended Questions in Dental Case Reports.

Journal of dentistry·2026

Same author

Consensus on the use of artificial intelligence in the management and measurement of vestibular schwannomas: A protocol for a modified delphi consensus.

Neuroradiology·2026

Same author

DeepPTMPred: a multi-modal deep learning framework for accurate prediction of protein post-translational modification sites.

Briefings in bioinformatics·2026

Same journal

Generative morphodynamic forecasting enables robust zero-shot volumetric medical segmentation.

Medical image analysis·2026

Same journal

ContiMorph: An unsupervised learning framework for cardiac motion tracking with time-continuous diffeomorphism.

Medical image analysis·2026

Same journal

MedP-CLIP: Medical CLIP with region-aware prompt integration.

Medical image analysis·2026

Same journal

Multi-organ guided diagnosis of mild cognitive impairment via hierarchical alignment and knowledge distillation.

Medical image analysis·2026

Same journal

SUDA: Simultaneous unsupervised knowledge distillation and adaptation of foundation models for efficient pathological image analysis.

Medical image analysis·2026

Same journal

Beyond the LUMIR challenge: The pathway to foundational registration models.

Medical image analysis·2026

See all related articles

This summary is machine-generated.

This study introduces a new deep learning method for medical image segmentation using sparse positive-only labels and out-of-distribution detection. It efficiently segments images without needing background annotations, improving accuracy and reliability.

Area of Science:

Computer Vision
Medical Imaging
Machine Learning

Background:

Medical image segmentation using deep neural networks faces challenges with time-consuming, expert-required pixel-level annotations.
Existing segmentation models struggle to detect out-of-distribution (OOD) pixels, leading to unreliable outputs in real-world applications.

Purpose of the Study:

To develop a novel segmentation approach addressing the limitations of sparse annotations and OOD detection in medical imaging.
To formulate multi-class segmentation with sparse positive-only annotations as a pixel-wise positive-unlabelled (PU) learning problem, leveraging OOD detection techniques.

Main Methods:

Proposed a framework utilizing positive-unlabelled (PU) learning and OOD detection techniques for segmentation.
Learned from sparsely annotated pixels of multiple positive-only classes, excluding background annotations, treating them as part of the OOD set.

Keywords:

Hyperspectral imaging One-class classification Out-of-distribution detection Positive-unlabelled learning Semantic segmentation Weakly supervised learning

Related Experiment Videos

Last Updated: Apr 9, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

Integrated pixel-level OOD detection approaches and proposed a cross-validation strategy treating held-out classes as OOD.

Main Results:

Demonstrated the framework's ability to segment medical images using only sparse positive-only annotations.
Showcased the integration of OOD detection for enhanced segmentation reliability.
Validated the robustness and generalization capability through extensive experiments on hyperspectral and RGB surgical imaging datasets.

Conclusions:

The proposed framework offers an effective solution for medical image segmentation with limited annotations.
Leveraging OOD detection techniques enhances the robustness of segmentation models.
The approach shows significant promise for improving segmentation accuracy and reliability in medical and surgical imaging.