Related Experiment Video
Updated: Jul 8, 2026
09:27
Measuring the Shape and Size of Activated Sludge Particles Immobilized in Agar with an Open Source Software Pipeline
Published on: January 30, 2019
7.4K
High-throughput and rapid classification on harmful algal bloom species based on mega image database and artificial
Junjie Zheng1, Zihan Sun1, Ruoyu Guo1
1Key Laboratory of Marine Ecosystem Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, 310012, China.
Marine Pollution Bulletin
|December 30, 2025
Summary
Artificial intelligence (AI) enhances microalgae identification, crucial for marine ecosystems and harmful algal blooms (HABs). AI models, particularly ViT, achieved high accuracy, aiding rapid detection and ecosystem protection.
Area of Science:
- Marine biology
- Computational biology
- Ecology
Background:
- Microalgae are vital to marine ecosystems and industry.
- Rapid identification of microalgae, especially harmful algal bloom (HAB) species, remains a significant challenge.
- Accurate identification is key for ecological monitoring and managing HABs.
Purpose of the Study:
- To develop and evaluate AI-based algorithms for accurate and rapid microalgae classification.
- To construct a comprehensive microalgae image database to train AI models.
- To assess the performance of different AI models, including DenseNet, EfficientNet, and Vision Transformer (ViT).
Main Methods:
- Creation of a diverse microalgae image dataset.
- Training and validation of deep learning models (DenseNet, EfficientNet, ViT).
- Comparative analysis of model performance and evaluation on field-collected samples.
Main Results:
- The Vision Transformer (ViT) model demonstrated superior classification performance.
- Dataset size and diversity were critical factors for improving AI model accuracy.
- AI model results showed strong agreement with traditional microscopic and metabarcoding analyses.
Conclusions:
- AI-powered microalgae classification offers a significant advancement in speed and accuracy.
- This technology provides a robust tool for early warning systems and rapid response to HABs.
- The findings support enhanced management and protection strategies for marine ecosystems through AI.
