|
Vol.15, No.3, August 2026. ISSN: 2217-8309 eISSN: 2217-8333
TEM Journal
TECHNOLOGY, EDUCATION, MANAGEMENT, INFORMATICS Association for Information Communication Technology Education and Science |
CNN Based Diatom Identification: A Systematic Review of Methods, Performance, and Educational Implications
Muhammad Iqbal Najib Fahmi, Susriyati Mahanal, Siti Zubaidah, Ibrohim Ibrohim
© 2026 Susriyati Mahanal, published by UIKTEN. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. (CC BY-NC-ND 4.0)
Citation Information: TEM Journal. Volume 15, Issue 3, Pages 3213-3227, ISSN 2217-8309, DOI: 10.18421/TEM153-99, August 2026.
Received: 27 December 2025.
Abstract:
Automated diatom identification using convolutional neural networks has attracted increasing attention for applications in environmental monitoring and microalgae learning, yet evidence remains fragmented across methodological and practical dimensions. The study conducted a systematic literature review of 27 articles related to diatom identification using convolutional neural networks. The review analyses dominant task formulations, model architectures, dataset characteristics, reported performance, technical challenges, and future research directions. The findings show that research is largely dominated by classification and detection tasks, relying on established architectures and microscopy derived datasets, while segmentation and regression-based approaches are comparatively limited. Although high performance is frequently reported, substantial heterogeneity in evaluation metrics, validation protocols, and dataset regimes constrains cross study comparability and limits generalizability. Persistent challenges include visual complexity, morphological similarity among taxa, class imbalance, limited interpretability, and reduced robustness beyond controlled settings. These issues are particularly critical for microalgae learning, where reliable and interpretable outputs are essential. Overall, this review highlights the need for data centred development, robustness-oriented evaluation, and educational alignment to advance scalable and pedagogically meaningful convolutional neural networks-based diatom identification systems.
Keywords – Convolutional neural networks, diatom identification, microalgae learning, systematic review. |
|
----------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------- |