An improved multivariate image analysis method for quality control of nanofiber membranes

Multivariate image analysis is a widely used technique for computing a spatial statistical characterization of an image. In this paper a modied method for multivariate image analysis is presented. The proposed method reformulates the approach previously presented by Bharati et al. [2004] extending its range of applicability by reducing its computational complexity and its memory requirements: this allows to take into consideration a larger set of spatial statistics to characterize the image texture. The proposed approach is applied to a case study concerning the estimation of the ber diameter distribution in nanostructured membranes. The results suggest that the optimum range of spatial statistics used for characterizing the image is related to the size of the main textural features.