Analysis of Interval Changes on Mammograms for Computer Aided Diagnosis

Abstract

The goal of this research is to develop a novel technique for computerized analysis of temporal differences between the most recent and previous mammograms of the same view in order to evaluate the usefulness of using interval change to distinguish between normal structures, benign masses, and malignant masses in computer-aided diagnosis (CAD). During this year, we have developed a multistage regional registration technique that combines global and local alignment procedures for identifying masses on temporal pairs of mammograms. In the first stage, the breast images from the current and previous mammograms were globally aligned by maximizing a mutual information measure. In the second stage a fan-shape search region based on polar coordinates was estimated on the previous mammogram. In the third stage a search for the best match between the lesion template from the current mammogram and a structure on the previous mammogram was carried out within the fan-shape region. We found that in 85% of the cases the corresponding regions on the previous and current mammogram were correctly identified. This technique can be useful for identification of corresponding lesions on temporal pairs of mammograms.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 1999
Accession Number
ADA381103

Entities

People

  • Lubomir X. Hadjiiski

Organizations

  • University of Michigan

Tags

DTIC Thesaurus Topics

  • Abstracts
  • Accuracy
  • Algorithms
  • Artificial Intelligence
  • Breast Cancer
  • Computer Vision
  • Computer-Aided Diagnosis
  • Computers
  • Databases
  • Detection
  • Identification
  • Image Processing
  • Materials
  • Neoplasms
  • Pattern Recognition
  • Standards
  • Two Dimensional

Fields of Study

  • Medicine
  • Physics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Oncology and Biomarker-Based Cancer Detection.