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 10, 2026

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

Uncertainty-aware multi-class brain tumor segmentation using Bayesian U-Net variants.

Rahul Pal¹, Sanoj Kumar², Gaurav Bhatnagar³

¹Department of Mathematics, UPES, Dehradun, India.

Biomedical Physics & Engineering Express

|April 8, 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 journal

Maintaining SUV accuracy in low-count PET with PETfectior: a deep learning denoising solution.

Biomedical physics & engineering express·2026

Same journal

Design and optimization of deep learning model based on multimodal data fusion for dynamic mental health assessment.

Biomedical physics & engineering express·2026

Same journal

Feasibility of a depth camera-based center of mass feedback system for closed-loop functional electrical stimulation therapy and visual feedback balance training.

Biomedical physics & engineering express·2026

Same journal

Simulation-driven deep learning for the diagnosis of middle ear pathologies using wideband acoustic immittance.

Biomedical physics & engineering express·2026

Same journal

Unsupervised machine learning-assisted multimodal characterization of cardiomyocytes on a thin-film-transistor microelectrode array (TFT-MEA).

Biomedical physics & engineering express·2026

Same journal

Assessment of experimental values of effective energies and beam quality correction factors for out-of-field dosimetry in external beam radiotherapy using radiophotoluminescent glass dosimeters.

Biomedical physics & engineering express·2026

This study introduces an uncertainty-aware framework for brain tumor segmentation using U-Net variants. It enhances model reliability by providing pixel-wise confidence maps, aiding clinical decision-making in magnetic resonance imaging (MRI).

Area of Science:

Medical Imaging
Artificial Intelligence

Background:

Accurate brain tumor segmentation in MRI is crucial for clinical applications.
Deep learning models show promise but lack reliable uncertainty estimation for clinical adoption.

Purpose of the Study:

To develop and evaluate an uncertainty-aware framework for multi-class brain tumor sub-region segmentation.
To investigate the impact of U-Net architectural variants on segmentation accuracy and reliability.

Main Methods:

Implemented an uncertainty-aware framework integrating multiple U-Net variants (Attention, Residual, Squeeze-Attention, CBAM).
Utilized Bayesian inference with Monte Carlo (MC) dropout for pixel-wise epistemic uncertainty estimation.
Evaluated on the BraTS 2020 dataset for whole tumor (WT), tumor core (TC), and enhancing tumor (ET) segmentation.

Keywords:

Bayesian deep learning Monte Carlo dropout U-Net variants attention mechanisms brain tumor segmentation medical image segmentation

Related Experiment Videos

Last Updated: Apr 10, 2026

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

Main Results:

U-Net variants achieved competitive segmentation accuracy for all tumor sub-regions.
Uncertainty maps highlighted distinct spatial patterns, especially at ambiguous tumor boundaries.
Uncertainty estimation provided complementary insights beyond standard accuracy metrics.

Conclusions:

The proposed framework enhances interpretability and reliability of automated brain tumor segmentation.
Uncertainty maps assist radiologists in identifying low-confidence regions, improving decision-making and trust.
Bayesian uncertainty estimation integrated with U-Net backbones offers clinically valuable confidence information.