Particle classification based on polarized light scattering and machine learning

Dongjian Zhan; Nan Zeng; Sirui Chen; Yonghong He; Hui Ma

doi:10.1117/12.2501071

25 October 2018 Particle classification based on polarized light scattering and machine learning

Dongjian Zhan, Nan Zeng, Sirui Chen, Yonghong He, Hui Ma

Proceedings Volume 10822, Real-time Photonic Measurements, Data Management, and Processing III; 108220J (2018) https://doi.org/10.1117/12.2501071
Event: SPIE/COS Photonics Asia, 2018, Beijing, China

Abstract

This paper focuses the data processing and then the multi-class classification of suspended particles using a new polarized light measurement scheme. Detection of multidimensional polarization parameters keeps the advantages of fast detection speed and non-invasive local analysis of light scattering method, and increases the information dimension of the analyzed particles. However, the polarization indices are numerous and interrelated. It is difficult to complete classification prediction by a few specific indices. More advanced algorithms are needed. In our research, we selected six kinds of representative particles and three typical machine learning algorithms. k-NN, Neural network and SVM methods were used to construct the classification models and solve different classification tasks. By comparison, we evaluated these models in terms of their performance for classification tasks in different aspects. Furthermore, we discuss how to improve the models by feature selection, and a rough prediction of the capability of each polarization index to reflect the particulate features was made.

Citation Download Citation

Dongjian Zhan, Nan Zeng, Sirui Chen, Yonghong He, and Hui Ma "Particle classification based on polarized light scattering and machine learning", Proc. SPIE 10822, Real-time Photonic Measurements, Data Management, and Processing III, 108220J (25 October 2018); https://doi.org/10.1117/12.2501071

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available