Learning-based automatic sensing and size classification of microparticles using smartphone holographic microscopy

Abstract

The accurate and fast size classification of microparticles is important in environmental monitoring and biomedical applications. Conventional methods for sensing and classifying microparticles require bulky optical setups and generally show medium performance. Accordingly, the development of a portable and smart platform for accurate particle size classification is essential. In this study, we propose a new sensing platform for automatic identification of microparticle types through the synergistic integration of smartphone-based digital in-line holographic microscopy (DIHM) and machine-learning algorithms. The smartphone-based DIHM system consists of a coherent laser beam, a pinhole, a sample holder, a three-dimensional printed attachment, and a modified built-in smartphone camera module. The portable device has a physical dimension of 4 × 8 × 10 cm³ and 220 g in weight. Holograms of various polystyrene microparticles with different sizes (d = 2–50 μm) were recorded with a wide field-of-view and high spatial resolution. To establish a proper classification model, tens of features including geometrical parameters and light-intensity distributions were extracted from holograms of individual particles, and five machine-learning algorithms were used. After examining the performance of several classifiers, the resulting support vector machine model trained by using three geometrical parameters and three extracted parameters from light-intensity distributions shows the highest accuracy in the particle classification of the training and test sets (>98%). Therefore, the developed handheld smartphone-based platform can be potentially utilized to cope with various imaging needs in mobile healthcare and environmental monitoring.