Where is the machine looking? Locating discriminative light-scattering features by class-activation mapping

Piedra, Patricio; Gobert, Christian; Kalume, Aimable; Pan, Yong‐Le; Kocifaj, Miroslav; Muinonen, K.; Penttilä, Antti; Zubko, Evgenij; Videen, Gorden

doi:10.1016/j.jqsrt.2020.106936

Cited by 11 publications

(5 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ideally, we control a trapped particle and actively rotate the particle over all positions while the ELS is being measured at each particle's position or orientation. Theoretically, scattering-pattern recognition can also be facilitated by using the method of machine learning, which is being explored in ELS of particles [216,217].…”

Section: Ot-elsmentioning

confidence: 99%

Optical trapping and laser-spectroscopy measurements of single particles in air: a review

Wang¹,

Pan²,

Videen³

2021

Meas. Sci. Technol.

Self Cite

View full text Add to dashboard Cite

From the early optical-tweezers approach, which uses a single tightly focused laser beam to levitate dielectric or absorbing micron-sized particles, to the recently developed optical traps such as the universal optical trap (UOT), which can trap particles of arbitrary chemical and physical properties in different media, optical trapping (OT) has evolved significantly over the last decades. Research in OT has been extended from single-particle control to single-particle measurements. One of the most rapid developments in OT is the combination of OT with advanced laser spectroscopic techniques to achieve on-trap single-particle studies. To date, a wide variety of single particles including carbons, dusts, metal oxides, pollens, spores, organic/inorganic droplets, etc have been stably trapped in air and characterized using Raman spectroscopy, cavity ringdown spectroscopy, light scattering, or laser-induced breakdown spectroscopy, etc. As single particles can be trapped stably in the UOT for long periods of time, temporal evolution of the chemical and physical properties of trapped particles can also be studied. Very recently, even chemical reactions of a single particle under controlled atmospheric environments have been investigated. This review updates the most recent developments in OT, with a particular emphasis on laser-spectroscopy measurements of single particles trapped in air.

show abstract

Section: Ot-elsmentioning

confidence: 99%

Optical trapping and laser-spectroscopy measurements of single particles in air: a review

Wang¹,

Pan²,

Videen³

2021

Meas. Sci. Technol.

Self Cite

View full text Add to dashboard Cite

show abstract

“…This dependency, especially for the orientation, since even the same particle can have an unlimited number of patterns, has been the limiting factor in using elastic scattered patterns as a major characterization method for unknown aerosol particles. Fortunately, recent seminal works by Piedra et al [ 125 , 126 ] applied machine learning (ML) algorithms to sets of computationally generated scattering patterns and achieved particle classification into groups, with an accuracy of ~70% for regularly shaped and ~90% for irregularly shaped particles. Ongoing efforts devoted toward advanced instrumentation for scattered light, improved and stable trapping techniques, as well as reliable ML algorithms, give hope to use elastic light scattering as a practical tool for fast characterization of aerosol particles.…”

Section: Single-particle Laser-based Characterizationmentioning

confidence: 99%

Optical-Trapping Laser Techniques for Characterizing Airborne Aerosol Particles and Its Application in Chemical Aerosol Study

2021

Self Cite

View full text Add to dashboard Cite

We present a broad assessment on the studies of optically-trapped single airborne aerosol particles, particularly chemical aerosol particles, using laser technologies. To date, extensive works have been conducted on ensembles of aerosols as well as on their analogous bulk samples, and a decent general description of airborne particles has been drawn and accepted. However, substantial discrepancies between observed and expected aerosols behavior have been reported. To fill this gap, single-particle investigation has proved to be a unique intersection leading to a clear representation of microproperties and size-dependent comportment affecting the overall aerosol behavior, under various environmental conditions. In order to achieve this objective, optical-trapping technologies allow holding and manipulating a single aerosol particle, while offering significant advantages such as contactless handling, free from sample collection and preparation, prevention of contamination, versatility to any type of aerosol, and flexibility to accommodation of various analytical systems. We review spectroscopic methods that are based on the light-particle interaction, including elastic light scattering, light absorption (cavity ring-down and photoacoustic spectroscopies), inelastic light scattering and emission (Raman, laser-induced breakdown, and laser-induced fluorescence spectroscopies), and digital holography. Laser technologies offer several benefits such as high speed, high selectivity, high accuracy, and the ability to perform in real-time, in situ. This review, in particular, discusses each method, highlights the advantages and limitations, early breakthroughs, and recent progresses that have contributed to a better understanding of single particles and particle ensembles in general.

show abstract

“…If a particle is known to be a uniform sphere, e.g., a water droplet, then it is also possible to measure the particle’s refractive index 20 – 22 . And, the recent integration of machine learning methods suggests the possibility of classifying the shapes of nonspherical particles using features in scattering patterns 23 , 24 .…”

Section: Introductionmentioning

confidence: 99%

The color of aerosol particles

Giri

Berg

2023

Sci Rep

View full text Add to dashboard Cite

Digital in-line holography (DIH) is an established method to image small particles in a manner where image reconstruction is performed computationally post-measurement. This ability renders it ideal for aerosol characterization, where particle collection or confinement is often difficult, if not impossible. Conventional DIH provides a gray-scale image akin to a particle’s silhouette, and while it gives the particle size and shape, there is little information about the particle material. Based on the recognition that the spectral reflectance of a surface is partly determined by the material, we demonstrate a method to image free-flowing particles with DIH in color with the eventual aim to differentiate materials based on the observed color. Holograms formed by the weak backscattered light from individual particles illuminated by red, green, and blue lasers are recorded by a color sensor. Images are reconstructed from the holograms and then layered to form a color image, the color content of which is quantified by chromaticity analysis to establish a representative signature. A variety of mineral dust aerosols are studied where the different signatures suggest the possibility to differentiate particle material. The ability of the method to resolve the inhomogeneous composition within a single particle in some cases is shown as well.

show abstract

Where is the machine looking? Locating discriminative light-scattering features by class-activation mapping

Cited by 11 publications

References 30 publications

Optical trapping and laser-spectroscopy measurements of single particles in air: a review

Optical trapping and laser-spectroscopy measurements of single particles in air: a review

Optical-Trapping Laser Techniques for Characterizing Airborne Aerosol Particles and Its Application in Chemical Aerosol Study

The color of aerosol particles

Contact Info

Product

Resources

About