Search

Image segmentations based on maximum likelihood or maximum a posteriori analyses of object textures usually assume parametric models (e.g., Gaussian) for distributions of these features. For real images, parameter accuracy and model stationarity may be elusive, so that model-free inference methods ought to have an advantage over those that are model-dependent. Functions of the relative entropy (RE) from information theory can produce minimum error, model-free inferences, and can detect the boundary of an image object by maximizing the RE between the pixel distributions inside and outside a flexible curve contour. A generalization of the RE -- the Jensen-Rényi divergence (JRD) -- computes optimal n-way decisions and can contour multiple objects in an image simultaneously. Seed regions expand naturally and multiple contours tend not to overlap. An edge detector based on the JRD, combined with multivariate pixel segmentation, generally improved the error of the segmentation. We apply these functions to contour patient anatomy in X-ray computed tomography for radiotherapy treatment planning.

Autoradiography can generate large quantities of information related to brain metabolism, blood flow, transport across the blood-brain barrier, neurotransmitter-receptor binding and other aspects of brain function. Three-dimensional (3D) reconstruction of digitized autoradiograms provides a mechanism for efficient analysis of function, in detail, over the entire brain. 3D reconstructions of the mean and variance can be obtained by superimposing data from similar experiments, leading ultimately to 3D reconstructions of differences with statistical tests of significance. Image registration is essential for reconstruction, and this article reports two independent algorithms for coronal image alignment that have been successfully implemented in computer programs. The first algorithm superimposes the centroids and principal axes of serial images; the extent and direction of the translation and rotation required for each image is obtained from an analysis of the inertia matrix of that image. The second algorithm matches the edges of structure features in serial-adjacent images, from analyses of the cross-correlation function of each pair of adjacent images. The cross-correlation method requires a great deal more computation than the principal axes method, but it can align damaged sections not reliably treated by the principal axes method. The methods are described in detail, and a quantitative assessment of the registration of non-identical images is considered.