Events Calendar

Non-Gaussian Image Structure, and the Rationale for Tomographic Image Reconstruction

Speaker: Dr. Craig Abbey, UC Santa Barbara

Talk Title: Non-Gaussian Image Structure, and the Rationale for Tomographic Image Reconstruction

Series: Medical Imaging Seminar Series

Abstract: Gaussian stochastic processes are a common model for medical and scientific images, leading to Gaussian statistical properties characterized by mean, variance, and correlations. However it is clear that real image ensembles have higher-order non-Gaussian structures that are not fully described by these statistics. Furthermore, studies in vision science suggest that he visual system is tuned to these non-Gaussian components.

This seminar will present one way to quantify non-Gaussian statistical structure in images, called Laplacian fractional entropy (LFE). We will then see some ways that LFE is influenced by factors of interest in medical imaging, such as image processing of full-field digital mammograms or different breast imaging technologies. Finally, we will consider the rationale for tomographic reconstruction of projection images, using LFE to address the possibility that the primary benefit of image reconstruction is to repackage the statistical properties of the acquired data.

Biography: Craig K. Abbey received his PhD in applied mathematics from the University of Arizona in 1998. He was a postdoctoral fellow in medical physics at Cedars-Sinai Medical Center and UCLA from 1998 to 2001. From 2001 to 2004 he was a member of the faculty in biomedical engineering at UC Davis, where he retains an adjunct position. His primary affiliation is in the Dept. of Psychological and Brain Science at UC Santa Barbara. His research focuses on how useful information is extracted from images in the presence of noise and other signal distortions. Methods for investigating this topic include theoretical analysis of image statistics as well as visual psychophysics for evaluating human observer performance.


Host: Professor Krishna Nayak

More Info: http://mhi.usc.edu/medical-imaging-seminar-series/

Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248

Audiences: Everyone Is Invited

Contact: Talyia Veal

Event Link: http://mhi.usc.edu/medical-imaging-seminar-series/