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

Representation learning for mammography mass lesion classification with convolutional neural networks.

John Arevalo1, Fabio A González1, Raúl Ramos-Pollán2

  • 1Universidad Nacional de Colombia, Bogotá, Colombia.

Computer Methods and Programs in Biomedicine
|February 1, 2016
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Interpretable Whole-Breast Radiomic Biomarkers for Exploratory Assessment of HER2 + Breast Cancer in Digital Mammography.

Journal of imaging informatics in medicine·2026
Same author

A Risk Analysis Tool for Medical Studies.

Studies in health technology and informatics·2026
Same author

Towards deep-learning based detection and quantification of intestinal metaplasia on digitized gastric biopsies: a multi-expert comparative study.

Scientific reports·2026
Same author

Interpretable weakly-supervised learning through kernel density matrices: A digital pathology use case.

PloS one·2025
Same author

Alternate dyes for image-based profiling assays.

SLAS discovery : advancing life sciences R & D·2025
Same author

Morphological map of under- and overexpression of genes in human cells.

Nature methods·2025

This study introduces a deep learning framework for automated breast cancer diagnosis from mammograms, improving lesion classification accuracy. The novel approach enhances diagnostic performance compared to traditional methods.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Oncology

Background:

  • Automated classification of breast lesions in mammography remains a challenge.
  • Current methods often rely on hand-crafted features, limiting performance.
  • Deep learning offers potential for automated feature learning in medical diagnostics.

Purpose of the Study:

  • To develop an innovative representation learning framework for breast cancer diagnosis using mammography.
  • To integrate deep learning techniques for automatic feature extraction, bypassing manual feature design.
  • To improve the accuracy of breast mass lesion classification.

Main Methods:

  • A new biopsy-proven dataset of 736 mammograms from 344 patients was curated.
  • A two-stage method involving image preprocessing and supervised deep learning was employed.
Keywords:
Breast cancerComputer-aided diagnosisConvolutional neural networksFeature learningMammography

Related Experiment Videos

  • Convolutional neural networks were used for supervised representation learning, differing from traditional descriptor-based approaches.

    • Main Results:

    • The developed deep learning framework achieved an Area Under the ROC Curve (AUC) of 0.822, outperforming traditional methods like HOG and HGD (AUC 0.787).
    • The model surpassed hand-crafted features that utilized radiologist segmentation data.
    • Combining learned and hand-crafted features yielded the highest AUC score of 0.826 for mass lesion classification.

    Conclusions:

    • A novel deep learning framework for automated classification of breast mass lesions in mammography was successfully developed.
    • The proposed method demonstrates superior performance in breast cancer diagnosis.
    • This framework represents a significant advancement in automated mammographic analysis.