Machine-learning approach to holographic particle characterization

Abstract

Holograms of colloidal dispersions encode comprehensive information about individual particles' three-dimensional positions, sizes and optical properties. Extracting that information typically is computationally intensive, and thus slow. Here, we demonstrate that machine-learning techniques based on support vector machines (SVMs) can analyze holographic video microscopy data in real time on low-power computers. The resulting stream of precise particle-resolved tracking and characterization data provides unparalleled insights into the composition and dynamics of colloidal dispersions and enables applications ranging from basic research to process control and quality assurance.

Holographic Characterization of Protein Aggregates

“…14,15,25 Each fit can be performed in a few tens of milliseconds using automated feature detection 27 and image recognition algorithms. 28 A single fit suffices to characterize a single protein aggregate.…”

Holographic characterization of colloidal fractal aggregates

“…Features associated with dispersed particles are identified in digitized holographic images 16 and are analyzed with Eqs. (1) and (2) using methods that have been described in detail elsewhere 2,3 . The example in Figure 1(b) is a 201 pixel × 201 pixel region of interest cropped from the normalized hologram, b ( r ), obtained from I ( r ) according to Eq.…”

“…A single snapshot of an individual colloidal particle can be analyzed in several milliseconds using standard computer hardware 3,17 . Characterization data therefore can be acquired in real time as particles flow down the microfluidic channel 17 .…”

“…Characterizing a single sphere requires a few milliseconds on a standard computer 3 , which is fast enough that several thousand spheres can be characterized in under ten minutes. Analyzing holograms of aspherical particles and colloidal clusters is substantially more challenging 4 , particularly if no information is available a priori about the particles’ geometry.…”

“…Similar detection and characterization challenges arise in the precision slurries used by the semiconductor manufacturing for chemical-mechanical planarization 11 . As a particle-resolved measurement technique, holographic characterization naturally differentiates micrometer-scale particles by size and composition 3 . The effective-sphere model extends these capabilities to include assessment of particle morphology without sacrificing speed or ease of use.…”