IMPULSE RESPONSE CHARACTERIZATION OF BREAST TOMOSYNTHESIS RECONSTRUCTION WITH PARALLEL IMAGING CONFIGURATIONS

Author

Apuroop Balla, Southern Illinois University CarbondaleFollow

Date of Award

12-1-2010

Degree Name

Master of Science

Department

Electrical and Computer Engineering

First Advisor

Chen, Dr.Ada

Abstract

It's universally accepted that early detection, diagnosis, and suitable treatment are best hope to increase the survival chance for breast cancer (BC) patients. To date, the most cost effective method for breast screening and early detection is mammography. The mammography technique uses a low-dose X-ray system to acquire images of the breast. Limitations of mammography exist, including 20% false negative rate, many call backs for screening, low positive predictive value of about 15% to 34% from biopsy. Digital Breast tomosynthesis (DBT) is a new technique that acquires limited angle tomographic projection images to reconstruct three-dimensional information of the breast. The development of tomosynthesis is a substantial improvement over conventional tomography in that it allows retrospective reconstruction of an arbitrary number of planes with a acquisition dataset. In this project, we investigated three different tomosynthesis image reconstruction algorithms including shift-and-add (SAA), back projection (BP), filtered back projection (FBP) with parallel imaging geometry to improve breast cancer detection. In this project we also examined the effect of varying the number of projection images (N) and total view angle (VA) for each reconstruction algorithm characterized by impulse response (IR) analysis. IR data were generated by simulating the projection images of a very thin wire, using various combinations of VA and N. Based on this project, we hope to optimize the image configuration to contribute to the breast tomosynthesis field.