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

Search research articles

Related Experiment Video

Updated: Jan 15, 2026

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

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

Published on: November 30, 2022

CBVLM: Training-free explainable concept-based Large Vision Language Models for medical image classification.

Cristiano Patrício¹, Isabel Rio-Torto², Jaime S Cardoso³

¹Universidade da Beira Interior, Covilhã, Portugal; INESC TEC, Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal; NOVA LINCS, Portugal.

Computers in Biology and Medicine

|October 11, 2025

Summary

Related Concept Videos

Vision

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

A Literature Review on Example-Based Explanations in Medical Image Analysis.

Journal of healthcare informatics research·2026

Same author

Conditional Generative Adversarial Network for Predicting the Aesthetic Outcomes of Breast Cancer Treatment.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2025

Same author

H&E to IHC virtual staining methods in breast cancer: an overview and benchmarking.

NPJ digital medicine·2025

Same author

Bridging Domain Gaps in Computational Pathology: A Comparative Study of Adaptation Strategies.

Sensors (Basel, Switzerland)·2025

Same author

Causality, Machine Learning, and Feature Selection: A Survey.

Sensors (Basel, Switzerland)·2025

Same author

Racial Differences in Temporal Thermometry and Association with Delayed Sepsis Bundle Care.

medRxiv : the preprint server for health sciences·2025

Same journal

Rapid personalisation of cardiovascular models using invasively measured right ventricular pressure.

Computers in biology and medicine·2026

Same journal

Biologically inspired mechanisms for enhancing robustness in EEG signal modeling: Challenges, opportunities, and perspectives.

Computers in biology and medicine·2026

Same journal

Machine learning-based detection of missed inspiratory efforts using esophageal pressure during noisy pressure support ventilation.

Computers in biology and medicine·2026

Same journal

A computational model of chemically- and mechanically-induced thrombus formation in cerebral aneurysms.

Computers in biology and medicine·2026

Same journal

An improved catch fish optimization based deep learning model for Parkinson disease classification using EEG signal.

Computers in biology and medicine·2026

Same journal

Assessing the robustness of evaluation metrics for synthetic ECG signal quality.

Computers in biology and medicine·2026

See all related articles

This summary is machine-generated.

This study introduces CBVLM, a novel method using Large Vision-Language Models (LVLMs) to improve medical AI interpretability and reduce data annotation needs. CBVLM achieves superior performance over existing methods with minimal data.

Area of Science:

Artificial Intelligence
Medical Imaging
Machine Learning

Background:

Deep learning adoption in medicine faces challenges with annotated data availability and model interpretability.
Concept Bottleneck Models (CBMs) offer interpretability but increase annotation burden and require retraining for new concepts.

Purpose of the Study:

To propose CBVLM, a methodology leveraging Large Vision-Language Models (LVLMs) to address data annotation and interpretability challenges in medical AI.
To reduce annotation costs and improve explainability in medical image analysis.

Main Methods:

CBVLM prompts LVLMs to identify concepts in images and then classify images based on these concepts.
A retrieval module is integrated for efficient in-context learning, selecting optimal examples.

Keywords:

Concept Bottleneck Models Concept-based explanations Explainability In-context learning Large Vision-Language Models Medical imaging Multimodal Large Language Models

Related Experiment Videos

Last Updated: Jan 15, 2026

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

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

Published on: November 30, 2022

The approach grounds final diagnoses in predicted concepts for explainability.

Main Results:

CBVLM consistently outperforms traditional CBMs and supervised methods across four medical datasets.
The methodology demonstrates effectiveness using twelve different LVLMs, both generic and medical.
CBVLM requires no model training and minimal annotated examples, significantly lowering annotation costs.

Conclusions:

CBVLM offers a powerful, explainable, and data-efficient solution for medical AI.
Leveraging LVLMs' few-shot capabilities, CBVLM overcomes key limitations of current deep learning approaches in healthcare.