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

MS-HIENet: multi-scale hybrid implicit-explicit registration network.

Physics in medicine and biology·2026

Same author

Association Between Oral Microbiota and Human Brain Glioma Grade: A Case-Control Study.

Frontiers in microbiology·2021

Same author

Ferroptosis Mediated by Lipid Reactive Oxygen Species: A Possible Causal Link of Neuroinflammation to Neurological Disorders.

Oxidative medicine and cellular longevity·2021

Same author

Is Self-Control an "Angel" or a "Devil"? The Effect of Internet Game Disorder on Adolescent Subjective Well-Being.

Cyberpsychology, behavior and social networking·2021

Same author

Centrosomal-associated Proteins: Potential therapeutic targets for solid tumors?

Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie·2021

Same author

Evidence Mapping of Proton Therapy, Heavy Ion Therapy, and Helical Tomotherapy for Gastric Cancer.

Oncology research and treatment·2021

Same journal

Deep learning-based dose prediction to enhance planning efficiency in cervical brachytherapy with hybrid applicators.

Physics in medicine and biology·2026

Same journal

Corrigendum: Referenceless MR thermometry-a comparison of five methods (2017<i>Phys. Med. Biol</i>.<b>62</b>1-16).

Physics in medicine and biology·2026

Same journal

Corrigendum: Measured and Monte Carlo simulated electron backscatter to the monitor chamber for the varian TrueBeam linac (2016<i>Phys. Med. Biol</i>.<b>61</b>8779).

Physics in medicine and biology·2026

Same journal

Corrigendum: 3D range-modulator for scanned particle therapy: development, Monte Carlo simulations and experimental evaluation (2017<i>Phys. Med. Biol</i>.<b>62</b>7075).

Physics in medicine and biology·2026

Same journal

Recent progress in applications of computing to radiotherapy (ICCR 2016).

Physics in medicine and biology·2026

Same journal

Novel TMS coils designed using an inverse boundary element method.

Physics in medicine and biology·2026

See all related articles

Search research articles

Related Experiment Video

Updated: Jun 19, 2025

Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function

Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function

Published on: April 12, 2024

Hierarchical multi-level dynamic hyperparameter deformable image registration with convolutional neural network.

Zhenyu Zhu¹, Qianqian Li², Ying Wei^1,3

¹School of Control Science and Engineering, Shandong University, Jinan, People's Republic of China.

Physics in Medicine and Biology

|July 25, 2024

Summary

This summary is machine-generated.

This study introduces a dynamic hyperparameter block for deep learning deformable image registration (DLDIR), enabling fast hyperparameter selection and improving accuracy. The novel method significantly reduces training time and enhances registration performance on brain and lung datasets.

Keywords:

deformable image registration dynamic convolution feature statistics hierarchical multi-level architecture regularization hyperparameters

More Related Videos

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

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

Related Experiment Videos

Last Updated: Jun 19, 2025

Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function

Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function

Published on: April 12, 2024

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

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

Area of Science:

Medical Imaging
Artificial Intelligence
Computer Vision

Background:

Deep learning deformable image registration (DLDIR) requires extensive hyperparameter tuning, which is time-consuming and resource-intensive.
Current DLDIR methods often involve numerous independent experiments for regularization hyperparameter selection.
Improving registration accuracy and deformation field regularity are key challenges in DLDIR.

Purpose of the Study:

To develop a DLDIR method that allows for single training and fast regularization hyperparameter selection during inference.
To enhance registration accuracy and deformation field regularity.
To reduce the computational cost associated with hyperparameter tuning in DLDIR.

Main Methods:

Proposed a novel dynamic hyperparameter block incorporating a distributed mapping network, dynamic convolution, attention feature extraction, and instance normalization.
Encoded input features and regularization hyperparameters into learnable feature variables and dynamic convolution parameters.
Implemented a hierarchical multi-level architecture for the dynamic hyperparameter block, replacing single-level residual blocks.

Main Results:

Reduced the percentage of folding (|Jϕ|⩽0) by 28.01% and 9.78% on the OASIS dataset, improving Dice similarity coefficient by 1.17% compared to LapIRN and CIR.
Reduced folding by 10.00% and 5.70% on the DIR-Lab dataset, decreasing target registration error by 10.84% and 10.05% compared to LapIRN and CIR.
Demonstrated reduced training time and superior registration accuracy and deformation smoothness compared to state-of-the-art methods.

Conclusions:

The proposed method enables rapid registration deformation field generation for arbitrary hyperparameters during inference.
Achieved state-of-the-art performance in registration accuracy and deformation smoothness on benchmark datasets.
Offers a significant reduction in training time compared to traditional DLDIR approaches with fixed hyperparameters.