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

Magnetic Resonance Imaging

Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) is a noninvasive medical imaging technique based on a phenomenon of nuclear physics discovered in the 1930s, in which matter exposed to magnetic fields and radio waves was found to emit radio signals. In 1970, a physician and researcher named Raymond Damadian noticed that malignant (cancerous) tissue gave off different signals than normal body tissue. He applied for a patent for the first MRI scanning device in clinical use by the early 1980s. The early MRI...

You might also read

Related Articles

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

Sort by

Same author

Preserving the Modiolus: Enhanced Functional and Aesthetic Restoration with the Modified Bernard-Webster Technique in Lip Reconstruction.

Plastic and reconstructive surgery·2026

Same author

Duplex MIRA assays for simultaneous detection of novel goose astrovirus and goose circovirus.

BMC veterinary research·2026

Same author

Establishment of a Rapid <i>Escherichia coli</i> Detection Method Based on MIRA-<i>Pf</i>Ago.

Biosensors·2026

Same author

The ecnAB toxin-antitoxin system modulates avian pathogenic Escherichia coli virulence through regulating the capsular sialic acid biosynthesis pathway.

Veterinary microbiology·2026

Same author

The application of a nomogram model integrating clinical factors and multimodal MRI radiomics features for predicting cervical lymph node metastasis for patients with oral tongue squamous cell carcinoma in different tumor stages.

Journal of cranio-maxillo-facial surgery : official publication of the European Association for Cranio-Maxillo-Facial Surgery·2026

Same author

Automated detection and classification of maxillary sinus variations using slice-based and full-volume CBCT deep learning models.

BMC oral health·2026

Search research articles

Related Experiment Video

Updated: Jun 26, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Swin MoCo: Improving parotid gland MRI segmentation using contrastive learning.

Zi'an Xu¹, Yin Dai¹, Fayu Liu²

¹Northeastern University, Shenyang, China.

Medical Physics

|May 15, 2024

Summary

Swin MoCo, a novel contrastive learning network, enhances parotid gland segmentation on limited medical datasets. This method improves segmentation accuracy and is computationally efficient, offering a new approach for medical image analysis.

Keywords:

contrastive learning image segmentation parotid gland tumor transformer unsupervised learning

More Related Videos

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

Related Experiment Videos

Last Updated: Jun 26, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

Area of Science:

Medical Imaging
Deep Learning
Computer Vision

Background:

Accurate segmentation of parotid glands and tumors in MR images is crucial for effective treatment of parotid gland tumors.
Parotid gland segmentation presents challenges due to variable gland shapes and low contrast with surrounding tissues.
Limited availability of large, well-annotated medical image datasets hinders the progress of deep learning applications in this field.

Purpose of the Study:

To address the limitations of small datasets in medical image segmentation using unsupervised learning methods.
To leverage contrastive learning, a rapidly developing unsupervised technique, to improve parotid gland segmentation.
To explore the potential of contrastive learning for enhancing the performance of deep learning models on medical imaging tasks.

Main Methods:

Development of Swin MoCo, a momentum contrastive learning network utilizing the Swin Transformer architecture.
Employing ImageNet pre-trained weights for the Swin MoCo backbone to enhance training on smaller medical image datasets.
Utilizing transfer learning to fine-tune the Swin MoCo model for improved segmentation performance.

Main Results:

Swin MoCo achieved significant improvements in parotid gland segmentation, reaching 89.78% DSC and 90.08% mAcc.
The model demonstrated strong performance on the Synapse multi-organ CT dataset when used as a pre-trained model for Swin-Unet, outperforming existing results.
Achieved 79.66% DSC and 12.73 HD on the Synapse dataset, showcasing its effectiveness across different medical imaging tasks.

Conclusions:

Swin MoCo offers a computationally inexpensive solution, requiring only 4 hours of training on a single V100 GPU.
The proposed method provides a novel approach to enhance performance on tasks involving small medical image datasets.
The code for Swin MoCo is publicly available, facilitating further research and development in medical image segmentation.