Analysis and Reconstruction of Medical Images Using Prior Information

Valen Johnson, James Bowsher, Jiang Qian, Ronald Jaszczak, Carey Floyd
Duke University

Nov 30 1991

We propose a Bayesian model for medical image analysis that permits prior structural information to be incorporated into the estimation of image features. Although the proposed methodology can be applied to a broad spectrum of images, we restrict attention here to emission computer tomography (ECT) images, and in particular single photon emission computer tomography (SPECT) images.

Inclusion of prior information is accomplished using the image model described in Johnson (1991a). A distinguishing feature of this model is the specification of a hierarchical structure for image generation that explicitly incorporates region parameters. These parameters are the mechanism by which we incorporate prior information regarding image structures. Importantly, these region identifiers allow prior information to be incorporated in a non-deterministic fashion, thus permitting prior structural information to be modified by image data with minimal introduction of residual artifacts. Furthermore, the resulting statistical model permits formation of previously unidentified structures based on the observed data likelihood.


PDF icon 1992-18.pdf