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 Experiment Videos

Technical architecture for integrating vision-language models with DICOM viewers.

Marco Albera1,2, Anna Colarieti3,4, Irene Albera4

  • 1SCDU Radiodiagnostica, Ospedale Maggiore Della Carità, Novara, Italy. marco.albera@icloud.com.

European Radiology Experimental
|July 1, 2026
PubMed
Summary

Related Concept Videos

Vision01:24

Vision

Vision is the result of light being detected and transduced into neural signals by the retina of the eye. This information is then further analyzed and interpreted by the brain. First, light enters the front of the eye and is focused by the cornea and lens onto the retina—a thin sheet of neural tissue lining the back of the eye. Because of refraction through the convex lens of the eye, images are projected onto the retina upside-down and reversed.

You might also read

Related Articles

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

Sort by
Same author

Revealing artifacts in contrast-enhanced mammography: prevalence and implications for diagnostic accuracy.

Quantitative imaging in medicine and surgery·2026
Same author

Performance tests of the B-RAD survey meter and portable γ-spectrometer in a clinical environment.

Physica medica : PM : an international journal devoted to the applications of physics to medicine and biology : official journal of the Italian Association of Biomedical Physics (AIFB)·2025
Same author

Deep-learning prediction of breast cancer hormone receptor status from CEM: a preliminary study.

European radiology experimental·2025
Same author

Voices of the future: insights from a survey on radiology training experiences and needs.

European journal of radiology·2025
Same author

Correction: Is there a need for a CT scan of the pancreatic phase? A perfusion and simulation study of the pancreas, an HCC, and the kidney cortex.

RoFo : Fortschritte auf dem Gebiete der Rontgenstrahlen und der Nuklearmedizin·2025
Same author

Contrast-enhanced mammography versus breast MRI: a multidimensional cost-effectiveness analysis.

La Radiologia medica·2025

This study introduces a radiology viewer that seamlessly integrates vision-language models (VLMs) into clinical workflows by re-rendering images offscreen. This ensures context-preserving, reproducible evidence generation with millisecond latency.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence in Medicine
  • Radiology Informatics

Background:

  • Integrating vision-language models (VLMs) into clinical radiology requires exporting 2D images that preserve diagnostic context.
  • Current methods like screenshots discard crucial context, reduce reproducibility, and disrupt workflows.

Purpose of the Study:

  • To describe a novel radiology viewer architecture for seamless VLM integration.
  • To enable context-faithful, reproducible evidence generation within the diagnostic environment.

Main Methods:

  • Developed a viewer architecture that serializes viewer state into replayable descriptors.
  • Implemented offscreen re-rendering using shared in-memory voxel data.
  • Designed a split workspace with a diagnostic grid and integrated chat sidebar for evidence staging.
Keywords:
Artificial intelligenceComputer graphicsDICOMRadiology workflowVision-language models

Related Experiment Videos

Main Results:

  • Achieved perfect reproducibility with pixel-identical images across repeated captures.
  • Demonstrated millisecond-scale capture latency (4.49–16.69 ms/frame).
  • Utilized high-efficiency image containers (HEIC) for storage-efficient output (27.9 kB/frame).

Conclusions:

  • The proposed architecture supports reproducible, low-latency, and storage-efficient evidence assembly.
  • Enables seamless integration of VLMs into diagnostic radiology workflows.
  • Facilitates traceable, auditable, and reproducible evidence generation for AI-assisted diagnostics.