Shape and size determination by laser diffraction: Feasibility of data analysis by neural networks

The feasibility of the inversion of laser diffraction data for size and sha pe distribution by neural networks has been investigated by computer simula tion. Neural networks trained with diffraction patterns of elliptical parti cles with different sizes and aspect ratios (axis ratios) were able to reco ver simultaneously both the size and aspect ratio distributions in a few mi lliseconds on a common PC.