Few-View Tomographic Reconstruction of Technetium-99m-Sestamibi Distribution for the Detection and Differentation of Breast Lesions

Abstract

The broad objective of the research is to develop and evaluate methods for the reconstruction of dedicated breast single-photon emission computed tomography (SPECT) scintimammography (S%4) images from a relatively small number of projection views. We have pursued a strategy involving sinogram preprocessing, in which each projection view is first smoothed using Fourier or spline-based techniques and then additional projection views are interpolated, again using Fourier or spline- based techniques, prior to reconstruction by filtered backprojection (FBP). The spline-based projection smoothing technique is a novel application of roughness- penalized nonparametric regression using an explicit Poisson model and is found to be superior to traditional Fourier smoothing techniques. Likewise, the spline- based projection interpolation approach is found to be more accurate in practical situations than the Fourier-based approach. %Then applied in concert, the two spline-based techniques yield high-quality images from as few as 15 projection views. We have also developed two alternatives to reconstruction by FBP, one based on the direct Fourier method, the other on direct reconstruction from the coefficients of splines fit to the projection views.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1999
Accession Number
ADA381135

Entities

People

  • Chin-tu Chen
  • Patrick J. Lariviere

Organizations

  • University of Chicago

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Breast Cancer
  • Cancer
  • Computational Science
  • Convolution Integrals
  • Detection
  • Detectors
  • Diagnostic Imaging
  • Information Processing
  • Information Science
  • Mathematical Filters
  • Periodic Functions
  • Power Spectra
  • Signal Processing
  • Three Dimensional
  • Tomography
  • Two Dimensional
  • X-Ray Computed Tomography

Fields of Study

  • Physics

Readers

  • Approximation Theory.
  • Image Processing and Computer Vision.
  • Medical Imaging.