We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Machine Learning Powered Microalgae Classification by Use of Polarized Light Scattering Data
Summary
Researchers developed a machine learning framework using polarized light scattering data to classify 35 categories of marine microalgae, finding that non-linear support vector machines achieved identification accuracy above 80% for more than 10 algal categories.
Microalgae are widely distributed in the ocean, which greatly affects the ocean environment. In this work, a dataset is presented, including the polarized light scattering data of 35 categories of marine microalgae. To analyze the dataset, several machine learning algorithms are applied and compared, such as linear discrimination analysis (LDA) and two types of support vector machine (SVM). Results show that non-linear SVM performs the best among these algorithms. Then, two data preparation approaches for non-linear SVM are compared. Subsequently, more than 10 categories of microalgae out of the dataset can be identified with an accuracy greater than 0.80. The basis of the dataset is shown by finding the categories independent to each other. The discussions about the performance of different incident polarization of light gives some clues to design the optimal incident polarization of light for future instrumentation. With this proposed technique and the dataset, these microalgae can be well differentiated by polarized light scattering data.
Sign in to start a discussion.
More Papers Like This
Pulse Feature-Enhanced Classification of Microalgae and Cyanobacteria Using Polarized Light Scattering and Fluorescence Signals
Researchers used polarization-sensitive flow cytometry with enhanced pulse feature analysis to classify microalgae and cyanobacteria in mixed samples, training classifiers on optical signatures that distinguish cell types without staining. The method achieved high classification accuracy and offers potential for rapid, label-free phytoplankton monitoring in environmental water samples.
Machine learning for microalgae detection and utilization
This review assessed machine learning applications for microalgae detection, classification, and utilization in aquaculture and bioproduction, finding that deep learning approaches achieve the highest accuracy for species identification from microscopy images. The authors highlighted ML as an enabling technology for automating microalgae monitoring and optimizing production in industrial bioreactors.
Classification of Microplastic Particles in Water using Polarized Light Scattering and Machine Learning Methods
Researchers developed a reflection-based, in-situ classification method for microplastic particles in water using polarized light scattering combined with machine learning, successfully identifying colorless particles in the 50-300 micrometer range. The approach circumvents transmission-based interference problems and offers a pathway toward continuous, large-scale microplastic monitoring in aquatic environments.
Digital Image Identification of Plankton Using Regionprops and Bagging Decision Tree Algorithm
Researchers developed a digital image classification system using machine learning to identify and count plankton from microscopy images. The method reduced the time and subjectivity of manual identification while maintaining accuracy. Automated plankton identification could also be adapted to distinguish microplastics from biological particles in environmental water samples.
Application of a convolutional neural network for automated multiclass identification of field-collected microplastics and diatom algae from optical microscopy images
Researchers developed and evaluated a convolutional neural network model using transfer learning to automatically classify field-collected microplastics and diatom algae from optical microscopy images, using a dataset of real microplastics sampled from a freshwater reservoir. The model achieved automated multi-class identification, including detection of diatom frustules that survive hydrogen peroxide processing, addressing challenges posed by the lack of standardised microplastic analysis protocols.