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: May 1, 2026

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

A calibration-aware hierarchical CNN-SWIN fusion framework for robust Cross-Dataset brain MRI analysis.

Ayesha Younis¹, Li Qiang¹, Abdur Rehman²

¹School of Microelectronics, Tianjin University, Tianjin, China.

Frontiers in Neuroscience

|April 30, 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

Roberts syndrome presenting with bilateral Cryptophthalmos: a case report.

Oxford medical case reports·2026

Same author

Effectiveness of a Teaching Session on Interpreting Common Orthopaedic Radiographs at a Major Trauma Centre.

Cureus·2026

Same author

Efficacy and safety of rupatadine in allergic rhinitis: a systematic review and meta-analysis of randomized controlled trials.

Annals of medicine and surgery (2012)·2026

Same author

Improving adherence to venous thromboembolism prophylaxis guidelines in surgical inpatients: a closed-loop clinical audit in a resource-limited setting.

Annals of medicine and surgery (2012)·2026

Same author

Adjunctive Minocycline in Acute Ischemic Stroke: A Systematic Review and Meta-Analysis of Randomized Controlled Trials.

Brain and behavior·2026

Same author

<b>Reviving Plecoptera research in Pakistan after five decades: faunistic review and redescription of <i>Mesonemoura skardui</i> (Aubert, 1959)</b>.

Zootaxa·2026

This study introduces a new deep learning framework for brain MRI analysis that improves reliability across different datasets. The model enhances accuracy and confidence calibration, making it safer for clinical use.

Area of Science:

Neuroimaging and Medical Data Analysis
Artificial Intelligence in Healthcare
Machine Learning for Neuroscience

Background:

Deep learning models, including CNNs and transformers, are vital for brain MRI analysis but struggle with dataset shift.
CNNs offer robust feature extraction but lack global context, while transformers capture long-range dependencies but can be less robust and miscalibrated on heterogeneous data.

Purpose of the Study:

To develop a robust deep learning framework for brain MRI analysis that addresses dataset shift challenges.
To improve the reliability and calibration of deep learning models in cross-dataset scenarios for neuroscience and clinical research.

Main Methods:

Proposed a calibration-aware hierarchical CNN-Transformer fusion framework integrating a pretrained CNN backbone with a transformer branch.

Keywords:

CNN-Transformer fusion Cross-Dataset generalization brain MRI analysis calibration-aware learning hierarchical feature fusion probabilistic calibration

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Related Experiment Videos

Last Updated: May 1, 2026

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Employed scale-aligned fusion using cross-attention mechanisms to allow local features to query global representations.

Evaluated the framework using a strict Cross-Dataset protocol, training on one dataset and testing on a distinct one.

Main Results:

The fusion model achieved competitive classification performance with improved probabilistic calibration over CNN-only and transformer-only baselines.
Attained an average accuracy of 99.20% with significantly lower Expected Calibration Error (ECE), Brier score, and Negative Log-Likelihood.
Demonstrated superior performance compared to standalone Swin Transformer and ResNet50 models.

Conclusions:

Calibration-aware hierarchical CNN-Transformer fusion enhances predictive reliability and robustness in cross-dataset brain MRI analysis.
The method improves the alignment between predictive confidence and empirical correctness, supporting safer large-scale analysis of heterogeneous MRI data.
Findings have significant implications for multi-center neuroscience studies and trustworthy clinical decision support systems.