New white paper reveals how pattern recognition techniques accurately identify different particle types in a heterogeneous sample
A new white paper by particle imaging authority Lew Brown of Fluid Imaging Technologies reveals how applying pattern recognition techniques to the use of advanced imaging particle analysis technology permits different particles in a heterogeneous sample to be automatically detected, differentiated, characterised and identified in real time.
The unprecedented capability works effectively even when the particles share the same equivalent spherical diameter (ESD) but are shaped completely differently.
Featuring an in-depth primer on particle image understanding (PIU), the white paper introduces Brown's proposed classification of image understanding into four distinct levels, starting from determining a simple particle count at Level Zero.
At Level One, particles are sized based on the assumption that all particles must be spherical, typical of laser diffraction technologies.
At Level Two, the true shape of a particle can be accurately characterised without the limitations of ESD.
At Level Three, higher level morphological attributes are measured such as the circularity of a particle based on its actual perimeter, among other measurements that become possible as increasing amounts of reliable data permit more subtle differentiations to be automatically recognised.
An example of Level Three image understanding features actual screen shots of different types of algal cells in a water sample that were automatically identified as members of different cell classes using the imaging and statistical pattern recognition capabilities of the Flowcam particle imaging and analysis system.
The automated Flowcam combines high speed particle imaging with 26 different measurement parameters, real-time pattern recognition and sophisticated data sorting and filtering capabilities that provide highly reliable, statistically significant data.
