Computer-Aided Characterization of Breast Masses on Volumetric Ultrasound Images: An Adjunct to Mammography

Abstract

The purpose of this project is to develop computer vision techniques for the analysis of sonographic images of breast masses, and to combine computerized sonographic and mammographic analyses. The techniques developed in this project are aimed at providing a second opinion to the radiologists in the task of making a biopsy recommendation. In the second year of the project, we have (1) compared the accuracy of the classifier designed in the first year of this project to that of experienced radiologists; (2) conducted studies on the effect of the developed classifier on radiologists' characterization of breast masses on ultrasound images; and (3) investigated methods for combining computer classification methods based on ultrasound and mammogram images. Our results indicate that the accuracy of our computer classifier is similar to that of experienced breast radiologists on our data set. We have also shown that experienced radiologists can significantly (p<0.006) improve their mass characterization accuracy on sonograms when aided by our algorithm. Our results on combining computer classification methods based on ultrasound and mammogram images indicate that multi-modality computer-aided diagnosis may further improve the classification accuracy.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 01, 2004
Accession Number
ADA431333

Entities

People

  • Berkman Sahiner

Organizations

  • University of Michigan

Tags

Communities of Interest

  • Air Platforms
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Biomedical Research
  • Breast Cancer
  • Carcinoma
  • Computer Vision
  • Computer-Aided Diagnosis
  • Computers
  • Data Acquisition
  • Data Sets
  • Electronic Mail
  • Health Services
  • Information Science
  • Machine Learning
  • Medical Personnel
  • Three Dimensional
  • Two Dimensional
  • Ultrasounds

Fields of Study

  • Medicine
  • Physics

Readers

  • Computer Vision.
  • Instructional Design and Training Evaluation.
  • Women's Health and Cancer Risk Research: African American Women and Pregnancy Outcomes.

Technology Areas

  • AI & ML