An Overcomplete Enhancement of Digital Mammograms

Abstract

This report deals with a technique for contrast enhancement of mammographic images by means of image fusion. We constructed a multiscale directional derivative based redundant wavelet representation to serve as a framework for image enhancement without introducing artifacts that could distort the appearance of mammographic features. After the wavelet decomposition is carried out, the transform coefficients are separately processed for enhancement of microcalcifications, circumscribed masses, and stellate lesions, and then fused for reconstruction of an image with improved visibility of mammographic abnormalities. Since the processing for enhancement of selected features and fusion of the resultant images are accomplished within a single wavelet transform framework, the method is computationally efficient and flexible enough for incorporation of different enhancement algorithms and their independent optimization. In order to evaluate the performance of the developed technique, we used quantitative criteria for comparison with histogram equalization and unsharp masking. Our method outperformed the two enhancement approaches during tests on simulated phantoms embedded in noise and on mammographic feature phantoms blended into mammograms. In addition, no spurious artifacts were observed as a result of our redundant wavelet transform based contrast enhancement scheme.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1999
Accession Number
ADA385824

Entities

People

  • Fred Taylor
  • Iztok Koren

Organizations

  • University of Florida

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Algorithms
  • Biomedical Research
  • Breast Cancer
  • Detection
  • Detectors
  • Dynamic Range
  • Filters
  • Frequency Domain
  • Image Processing
  • Information Science
  • Matched Filters
  • Pattern Recognition
  • Probability
  • Probability Density Functions
  • Signal Processing
  • Two Dimensional
  • Wavelet Transforms

Fields of Study

  • Medicine
  • Physics

Readers

  • Computer Vision.
  • Image Processing and Computer Vision.
  • Oncology and Biomarker-Based Cancer Detection.