Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Improving Translational Accuracy02:07

Improving Translational Accuracy

13.9K
Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...
13.9K
Improving Translational Accuracy02:07

Improving Translational Accuracy

3.5K
3.5K

You might also read

Related Articles

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

Sort by
Same author

How to Safely Deliver Stereotactic Ablative Radiotherapy With High-Dose or Multi-agent Systemic Therapy.

Cureus·2025
Same author

Concurrent Radiation With Nivolumab and Chemotherapy for Nodal-Only Distant Metastatic Stage IV Esophageal Carcinoma.

Cureus·2025
Same author

On the use of principal component analysis method to optimize sphere packing algorithm for lattice radiotherapy of large/bulky unresectable tumor.

Medical physics·2025
Same author

The Mechanism and Rate-Determining Step of Catalytic Ammonia Oxidation on Pd(332) at High Temperatures.

ACS catalysis·2025
Same author

Reported Early Clinical Outcomes of Forward-Planned Multileaf Collimator-Based 3-Dimensional Conformal Spatially Fractionated Radiation Therapy Technique for Large and Bulky Tumors.

International journal of radiation oncology, biology, physics·2025
Same author

Response to "Comment on 'Microcanonical treatment of HCl dissociative chemisorption on Au(111): Reactive dampening through inefficient translational energy coupling and an active surface'" [J. Chem. Phys. 162, 087101 (2025)].

The Journal of chemical physics·2025

Related Experiment Video

Updated: Dec 28, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K

Improving accuracy and robustness of deep convolutional neural network based thoracic OAR segmentation.

Xue Feng1,2, Mark E Bernard3, Thomas Hunter3

  • 1Department of Biomedical Engineering, University of Virginia, Charlottesville, VA 22903, United States of America.

Physics in Medicine and Biology
|February 21, 2020
PubMed
Summary

A deep convolutional neural network (DCNN) trained on public data showed reduced performance on local computed tomography (CT) scans. Adding just 10 local cases significantly improved accuracy, demonstrating effective transfer learning for medical image segmentation.

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

704

Related Experiment Videos

Last Updated: Dec 28, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

704

Area of Science:

  • Medical imaging analysis
  • Artificial intelligence in radiology
  • Deep learning for segmentation

Background:

  • Deep convolutional neural networks (DCNNs) excel at medical image segmentation, particularly for organs-at-risk (OARs) in computed tomography (CT).
  • Generalizability of DCNN models across different datasets remains a challenge, often leading to performance degradation.
  • Systematic differences between datasets can impact model accuracy and robustness.

Purpose of the Study:

  • To evaluate the generalization capability of a DCNN model trained on public thoracic OAR segmentation data when applied to a local dataset.
  • To identify and understand the systematic differences causing performance drops in DCNN models across datasets.
  • To develop and validate an optimal strategy for improving DCNN model accuracy and robustness on new, local datasets.

Main Methods:

  • A pre-trained DCNN model for thoracic OAR segmentation was tested on a local dataset.
  • Systematic differences between the public and local datasets were analyzed.
  • Transfer learning strategies, incorporating varying numbers of local cases, were employed to retrain the DCNN model.
  • Performance was evaluated based on segmentation accuracy and robustness.

Main Results:

  • The DCNN model exhibited significantly worse performance on the local dataset compared to the public dataset.
  • A subtle organ shift, attributed to abdominal compression during CT acquisition, was identified as a key factor for performance decline.
  • Incorporating as few as 10 new cases from the local institution restored the model's performance to the original level.
  • Transfer learning accelerated training time, though it resulted in slightly reduced performance for heart segmentation.

Conclusions:

  • DCNN model performance is sensitive to dataset shifts, even subtle ones like organ positioning variations.
  • A small number of local data cases, combined with transfer learning, is an effective strategy to adapt DCNN models for improved generalization in medical image segmentation.
  • Transfer learning offers an efficient approach to enhance DCNN model accuracy and robustness for clinical applications.