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

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

KPNA2 Drives Immunosuppression in Ovarian Cancer via CCL2/CCR2-Dependent MDSC Recruitment.

Cancer science·2026

Same author

A survey in Liangshan, Sichuan showed higher human papillomavirus infection rate and cervical cancer risk for Yi Chinese compared to Han Chinese.

Gynecology and pelvic medicine·2026

Same author

TGF-β inhibition enhances anti-tumor immunity and sensitizes PD-1 blockade therapy of bone marrow-derived myofibroblasts-induced "immunotherapy resistance" tumor model.

Cancer immunology, immunotherapy : CII·2026

Same author

Lung scRNA-seq reveals chronic inflammation and emphysemous phenotype in mice with osteogenesis imperfecta.

Frontiers in genetics·2026

Same author

Severity-related association between sleep-disordered breathing and nocturnal enuresis in children and adolescents and adolescents: a systematic review and meta-analysis.

BMC pediatrics·2026

Same author

Med-R1: Reinforcement Learning for Generalizable Medical Reasoning in Vision-Language Models.

IEEE transactions on medical imaging·2026

Same journal

Physiology-guided Self-supervised Learning for Simultaneous Dual-Tracer PET Separation.

IEEE transactions on medical imaging·2026

Same journal

Informed-Exploration Reinforcement Learning for Automated Virtual Coronary Intervention Planning.

IEEE transactions on medical imaging·2026

Same journal

4D Reconstruction of Fetal Left Ventricle from Echocardiography via 2.5D Radial Segmentation and Graph-Fourier Reconstruction.

IEEE transactions on medical imaging·2026

Same journal

Generalised Medical Phrase Grounding.

IEEE transactions on medical imaging·2026

Same journal

EndoLRMGS: Combining Large Reconstruction Modelling and Gaussian Splatting for Complete Endoscopic Scene Reconstruction.

IEEE transactions on medical imaging·2026

Same journal

A Neural-Analytical Fusion Scatter Correction Method for Multi-Source CT Using Equivalent High-Order Scatter.

IEEE transactions on medical imaging·2026

See all related articles

Search research articles

Related Experiment Video

Updated: Sep 15, 2025

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

Debiasing Medical Knowledge for Prompting Universal Model in CT Image Segmentation.

Boxiang Yun, Shitian Zhao, Qingli Li

IEEE Transactions on Medical Imaging

|July 15, 2025

Summary

This summary is machine-generated.

This study introduces a Debiased Universal Model (DUM) to reduce biases in medical image segmentation. DUM uses instance-level context to improve accuracy and generalization for diverse medical cases.

More Related Videos

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images

Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images

Published on: February 18, 2015

Related Experiment Videos

Last Updated: Sep 15, 2025

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

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images

Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images

Published on: February 18, 2015

Area of Science:

Artificial Intelligence
Medical Imaging
Computer Vision

Background:

Large language models offer medical prior knowledge for Universal Models (UM) in medical image segmentation.
Text prompts in UMs provide comprehensive knowledge but can introduce biases, especially for heterogeneous organs or rare cancers.
Existing UMs may struggle with specific medical cases due to inherent biases in general knowledge.

Purpose of the Study:

To propose a Debiased Universal Model (DUM) that mitigates knowledge biases in text prompts for medical image segmentation.
To leverage instance-level context information to correct biases introduced by organ-level text prompts.
To enhance the generalization capabilities of Universal Models in diverse and challenging medical scenarios.

Main Methods:

Developed a causal graph framework to identify and mitigate biases from universal knowledge.
Extracted organ-level text prompts using language models and instance-level context prompts from visual features.
Designed a standard UM (SUM) and a biased UM, obtaining the debiased output by subtracting the biased UM's output from the SUM's.
Validated the approach on multiple large-scale, multi-center external datasets and internal tumor datasets.

Main Results:

The proposed DUM method effectively reduces potential biases in medical image segmentation.
Demonstrated enhanced generalization ability in handling diverse medical scenarios and heterogeneous organs.
Achieved a 4.16% improvement over popular universal models on the AbdomenAtlas dataset.
Experimental results confirm the method's strong generalizability and bias mitigation capabilities.

Conclusions:

The Debiased Universal Model (DUM) successfully addresses the challenge of bias in Universal Models for medical image segmentation.
Instance-level context information is crucial for correcting biases originating from general medical knowledge.
DUM offers a promising approach for improving the reliability and applicability of AI in complex medical imaging tasks.