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

Redefining lightweight vision models for healthcare AI.

Linus Lee1, Zhibin Feng1, Jen Hong Tan1

  • 1Data Science and Artificial Intelligence Lab, Singapore General Hospital, Singapore, Singapore.

Frontiers in Artificial Intelligence
|June 17, 2026
PubMed
Summary

Related Concept Videos

Issues And Trends In Healthcare Delivery System01:29

Issues And Trends In Healthcare Delivery System

The issues and trends in healthcare delivery are constantly changing. The COVID-19 pandemic is one recent issue that wreaked havoc on healthcare systems, causing a shortage of healthcare workers, high demand for medicines and supplies, and increased medical expenditure due to a lack of insurance. Other issues include rising healthcare costs and care fragmentation.
Cost Containment
Payment for healthcare services has historically promoted adoption of costly and often unnecessary or inefficient...
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

Surgical outcomes following spinal fusion for cervical versus noncervical fractures in ankylosing spondylitis and DISH.

North American Spine Society journal·2026
Same author

Is Parkinson Disease a Risk Factor After Total Elbow Arthroplasty? Midterm Survivorship in a Matched Cohort.

Journal of hand surgery global online·2026
Same author

Paget's Disease of the Spine as a Catalyst for Osteosarcoma: A Narrative Review.

Cureus·2026
Same author

Hypoplasia of the Atlas Does Not Correlate With Subaxial Canal Stenosis: A Retrospective Cohort Study.

Cureus·2026
Same author

Medium Term Outcomes of Single-Level Posterior Cervical Foraminotomy versus Single-Level Anterior Cervical Decompression and Fusion versus Single-Level Cervical Disc Replacement for Cervical Radiculopathy.

Spine·2026
Same author

Hip Fracture Risk Following Lumbar Fusion: A Retrospective Cohort Study of Construct Length and Age-Stratified Outcomes.

European journal of orthopaedic surgery & traumatology : orthopedie traumatologie·2026

We developed ultra-lightweight medical vision transformers, MedLiT-seed and MedLiT-nano, with 2.1M and 0.75M parameters respectively. These efficient models achieve competitive performance in medical image analysis, outperforming larger architectures.

Area of Science:

  • Artificial Intelligence
  • Computer Vision
  • Medical Imaging

Background:

  • Traditional medical vision models are parameter-heavy, prompting research into architectural efficiency.
  • Achieving high classification performance without sacrificing efficiency is a key challenge.

Purpose of the Study:

  • Introduce MedLiT-seed (2.1M parameters) and MedLiT-nano (0.75M parameters), ultra-lightweight vision transformers.
  • Enable efficient and scalable medical image analysis.

Main Methods:

  • Utilized a streamlined Mixture-of-Experts (MoE) architecture with SwiGLU, grouped query attention, and depth-wise scaling.
  • Pre-trained models using masked autoencoding on ImageNet and MedMNIST, followed by fine-tuning on 12 MedMNIST 2D subsets.
  • Compared performance against benchmark models like ResNet, MedViT, and AutoML systems.
Keywords:
Grouped Query Attention (GQA)ImageNet datasetMedMNIST datasetMixture-of-Experts (MoE)SwiGLU feedforward networkVision Transformer (ViT)lightweight modelmasked AutoEncoder

Related Experiment Videos

Main Results:

  • MedLiT-seed achieved top AUC on 4 subsets and strong performance on others, outperforming models 10-20x larger.
  • MedLiT-nano matched or exceeded ResNet-18 and AutoML baselines on several subsets.
  • Transfer learning from ImageNet improved convergence and generalization; increasing embedding size was more impactful than increasing expert count.

Conclusions:

  • MedLiT's MoE-based token routing offers a viable pathway for competitive accuracy with minimal parameters (around 2M).
  • Selective computation routing through specialized experts is an effective design for compact medical vision models.
  • This architecture is suitable for low-resource settings and scalable fine-tuning, though multi-label task limitations require future refinement.