Wavelet Representations for Digital Mammography.

Abstract

This report describes recent progress in the development of a methodology for accomplishing adaptive contrast enhancement by multiscale representations. Our studies demonstrate that features extracted from multiresolution representations can provide an adaptive mechanism for the local emphasis of salient and subtle features of importance to mammography. We establish connections between enhancement techniques based on dyadic wavelet analysis and traditional unsharp masking and prove that two cases of linear enhancement are mathematically equivalent to traditional unsharp masking with Gaussian low-pass filtering. In addition, a digital image editor is described that allows radiologists to interactively indicate on computer screens (simultaneously displaying four views) regions diagnosed as probable cancer. Construction of average interpolation wavelets and Deslauriers-Dubuc representations on an interval provide radiologist with an interactive capability to process only suspicious regions and significantly reduce execution time. These representations (of arbitrary shape) provide improved sensitivity for the local emphasis of such features of importance to mammography including masses. spicules and microcalcifications. Finally, we report on the development of objective ways to assess the performance of wavelet image processing algorithms. Our objectives are to develop techniques to evaluate wavelet algorithms so they can then be optimized for clinical use in mammography.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 15, 1994
Accession Number
ADA292534

Entities

People

  • Andrew F. Laine

Organizations

  • University of Florida

Tags

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Breast Cancer
  • Computations
  • Detection
  • Detectors
  • Diagnostic Imaging
  • Digital Images
  • Frequency Response
  • Health Services
  • Image Processing
  • Lists (Data Structures)
  • Mammography
  • Mathematical Analysis
  • Mathematical Models
  • Medical Personnel
  • Two Dimensional

Readers

  • Computer Vision.
  • Oncology and Biomarker-Based Cancer Detection.
  • Theoretical Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference