Innovative software will be developed for geometrical analysis of three- dimensional (3D) computer reconstructions, or 3D models, derived from serial section data such as laser-scanning confocal microscope (LSCM), computed tomography and magnetic resonance image data. The new software will quantify numbers and sizes (surface area and volume) of individual 3D objects in a model, calculate shape descriptors of 3D objects, including branching structures, and measure distance relationships between objects in two different 3D models. It will require the development, implementation and testing of algorithms for rapid quantification of the geometry of 3D objects, determining 3D skeletons of objects and for measuring distances between objects and between points in space and 3D objects. Proof of concept and implementation will be tested on two experimental models: (1) Spatial relationships between nerve cells and their associated terminal synaptic fibers. (2) Characterization of the 3D branching geometry of sensory and sudomotor nerves in skin of normal and diabetic human. Synthesis of the software and measured results suggest the feasibility of the approach. Applications for the technology exist in both research and clinical diagnosis; for example, quantitative characterization of chemical neural pathways in brain and simplified quantitative testing for peripheral neuropathies with skin biopsy.
The proposed research will lead to software tools that will complement and build upon our extant 3D imaging, reconstruction and visualization software, ImageVolumes. ImageVolumes is being sold commercially for use in fields such as medical imaging, confocal microscopy, electron microscopy and industrial CT imaging. The addition of 3D geometrical analysis will greatly enhance the utility of the ImageVolumes product.
