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

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...
Improving Translational Accuracy02:07

Improving Translational Accuracy

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...

You might also read

Related Articles

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

Sort by
Same author

The Risks and Benefits of Synthetic Medical Images.

Radiology·2026
Same author

Deep-learning based quantitative evaluation of postoperative atelectasis following right upper lobectomy.

NPJ digital medicine·2026
Same author

Merlin: a computed tomography vision-language foundation model and dataset.

Nature·2026
Same author

The Effect of AI on the Radiologist Workforce: A Task-Based Analysis.

medRxiv : the preprint server for health sciences·2026
Same author

Improving Performance, Robustness, and Fairness of Radiographic AI Models with Finely-Controllable Synthetic Data.

Research square·2025
Same author

Leveraging large language models to extract smoking history from clinical notes for lung cancer surveillance.

NPJ digital medicine·2025

Related Experiment Video

Updated: Jul 2, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Improving language models for radiology speech recognition.

John M Paulett1, Curtis P Langlotz

  • 1School of Engineering and Applied Science, University of Pennsylvania, USA.

Journal of Biomedical Informatics
|September 2, 2008
PubMed
Summary

Speech recognition systems in radiology can be improved. Tailoring language models to body site and imaging modality significantly enhances word recognition accuracy, boosting radiologist productivity.

Related Experiment Videos

Last Updated: Jul 2, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Area of Science:

  • Medical Informatics
  • Natural Language Processing
  • Radiology

Background:

  • Speech recognition systems are widely used for radiology report generation due to efficiency and cost benefits.
  • Suboptimal accuracy in these systems can hinder radiologist productivity.
  • Analyzing large datasets of radiology reports is crucial for understanding system performance.

Purpose of the Study:

  • To identify key determinants of word frequency in radiology reports.
  • To assess the impact of speaker identity, body site, and imaging modality on speech recognition accuracy.
  • To provide recommendations for improving speech recognition system performance in radiology.

Main Methods:

  • Analysis of a de-identified database containing over two million radiology reports.
  • Statistical analysis to determine the influence of various factors on word and phrase frequency.
  • Comparison of the impact of speaker identity, body site, and imaging modality.

Main Results:

  • Body site and imaging modality were found to be strong determinants of word frequency, comparable to speaker identity.
  • Analysis revealed patterns in word and three-word phrase frequency influenced by anatomical location and imaging technique.
  • The findings highlight significant variations in language use based on clinical context.

Conclusions:

  • Speech recognition accuracy in radiology can be substantially improved by customizing language models.
  • Tailoring models to specific body sites and imaging modalities, which are known at report creation, is a promising strategy.
  • Enhanced speech recognition performance can lead to increased radiologist efficiency and cost savings.