You may download a PDF copy of this article from here.
This work was done for a defense lab involved in the development of computing solutions for numerical analysis and their applications, which include scientific visualization packages. The client required modules for visualizing tensor data. Tensors fields are generated in a broad range of areas like medical imaging, material sciences, electromagnetic etc. and hence require different visualization techniques based on the application area. Tensor data also tends to be large owing to 9 component values per cell for a 2nd order tensor.
With experience in writing the in-house scientific visualization package, ViewZ™, Zeus Numerix was able to provide modules for visualizing 2nd order tensors. These included hyper-streamlines, tensor ellipsoids and surface deformation. The modules were written in C++ and integrated with the existing codebase with the client. Python bindings were also provided for the same.
With the tensor modules integrated into client’s visualization system, one of the most complex parts was taken care.