Dimensionality Reduction for Classification of Stochastic Fibre Radiographs

Dimensionality reduction helps to identify small numbers of essential features of stochastic fibre networks for classification of image pixel density datasets from experimental radiographic measurements of commercial samples and simulations. Typical commercial macro-fibre networks use finite length fibres suspended in a fluid from which they are continuously deposited onto a moving bed to make a continuous web; the fibres can cluster to differing degrees, primarily depending on the fluid turbulence, fibre dimensions and flexibility. Here we use information geometry of trivariate Gaussian spatial distributions of pixel density among first and second neighbours to reveal features related to sizes and density of fibre clusters.