Fourier analysis and automated measurement of cell and fiber angular orientation distributions

This paper studies the application of the discrete Fourier transform (DFT) to predict angular orientation distributions from images of fibers and cells. Angular distributions of fibers in composites define their material properties. In biological tissues, cell and fiber orientation distributions are important since they define their mechanical properties and function. We developed a filtering scheme for the DFT to predict angular distributions accurately. The errors involved in this DFT technique and their sources were quantified through Monte Carlo simulation of computer-generated images. The knowledge of these errors allows one to verify the suitability of the method for a particular application. We found that the DFT method is most accurate for slender fibers, and propose a means to minimize errors by optimizing parameters. This method was applied to predict orientation distribution of cells and actin fibers in bio-artificial tissue constructs.