Geometry-driven visualization of microscopic structures in biology

Abstract

There are natural geometric patterns in biology. Tissue layers, for example, differ mainly in the spatial distributions, size and packing of microstructure components such as the red blood cells, nuclei and cytoplasm etc. Expressive visualization by using the N-point correlation functions, involves the discovery of feature spaces that estimate and spatially delineate component distributions unique to a salient tissue. These functions provide feature spaces that are used to set useful transfer functions. We obtain insightful 3D visualizations of the epithelial cell lining in mouse mammary ducts and evolving structures in a zebrafish embryo. These are large datasets acquired from light and confocal microscopy scanners respectively.