Deep learning-driven classification of 46 medicinal macrofungi species: comparative analysis, fusion strategies, and explainability insights.
7 affiliations
- 1Department of Biology, Faculty of Engineering and Natural Sciences, Osmaniye Korkut Ata University, 80000, Osmaniye, Turkey.
- 2Department of Nutrition and Dietetics, Faculty of Health Sciences, İstanbul Kültür University, 34191, Istanbul, Turkey.
- 3Institute of Artificial Intelligence, Ankara University, 06100, Ankara, Turkey.
- 4Graduate School of Natural and Applied Sciences, Ankara University, 06830, Ankara, Turkey.
- 5Department of Plant Food Products and Biofortification, Institute of Food Biotechnology and Genomics, National Academy of Sciences of Ukraine, Kyiv, 04123, Ukraine. [email protected].
- 6Department of Computer Engineering, Faculty of Engineering, Ankara University, 06830, Ankara, Turkey.
- 7Department of Biology, Faculty of Science, Ankara University, 06100, Ankara, Turkey. [email protected].
Abstract
Accurate identification of medicinal and wild macrofungi remains a major challenge due to high intra-species variability, visually overlapping morphologies, and inconsistent field-level imaging conditions. Existing deep learning studies typically focus on limited taxonomic scopes (5-12 classes) and lack systematic comparisons across modern convolutional and transformer architectures, leaving a significant gap in fine-grained fungal classification. This study introduces a comprehensive evaluation framework covering 46 macrofungal species, integrating 18 single backbones and three feature-level fusion strategies under a unified, fully reproducible protocol. All models were trained with identical preprocessing, stratified partitions, deterministic seeds, and a two-stage fine-tuning pipeline to ensure fair architectural comparison. Among single models, ConvNeXt-Base achieved the strongest performance with 96.59% accuracy, 0.966 macro-F1, 0.965 MCC, 0.9997 macro-AUC, and 0.1054 log-loss, exhibiting a near-diagonal confusion matrix where 25 classes reached perfect recall. Although feature-level fusion did not exceed the best single model in overall accuracy, it improved stability for several visually difficult taxa. The best ensemble, ConvNeXt-Base + DeiT-Base, achieved 95.85% accuracy, 0.958 macro-F1, and 0.9995 macro-AUC, remaining close to the ConvNeXt-Base baseline while providing complementary class-level robustness. The reported latency of ≈0.0021 s/image refers only to the precomputed-feature mode, in which backbone features are already cached and only the shared MLP head is evaluated. Therefore, this value should not be interpreted as full end-to-end inference latency for newly acquired field images. Explainability analyses using Grad-CAM and Integrated Gradients demonstrated that both single and fused models rely on biologically meaningful structures such as gill edges, pore surfaces, cap margins, and hymenophore textures. Overall, this study provides one of the most extensive quantitative benchmarks for macrofungal recognition to date and establishes a reproducible, architecture-agnostic foundation for future multimodal, large-scale, and explainable AI systems in fungal taxonomy.
Related
- Comparative mitochondrial genomics of the medicinal mushroom Phellinus tuberculosus reveals conserved core genes and dynamic genome evolution within Hymenochaetaceae
- When Clinical and Metabolomics Data Work Together: A Comparative Framework for Multimodal Disease Classification across Machine and Deep Learning.
- Comparative analysis of precursor supply and pathway reinforcement strategies for shikimic acid biosynthesis in Pseudomonas putida.
- Explainability from Training with Applications to TCR-Epitope Prediction
- Editorial: Biosynthetic pathways and metabolic engineering of bioactive natural products from medicinal plants
BibTeX
@article{sevindik2026deep,
title = {{Deep learning-driven classification of 46 medicinal macrofungi species: comparative analysis, fusion strategies, and explainability insights.}},
author = {Mustafa Sevindik and Aras Fahrettin Korkmaz and Fatih Ekinci and Eda Kumru and Tetiana Krupodorova and Ömer Burak Altindal and Mehmet Serdar Güzel and Ilgaz Akata},
journal = {BMC bioinformatics},
year = {2026},
month = aug,
doi = {10.1186/s12859-026-06584-z},
url = {https://europepmc.org/article/MED/42967984},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Deep learning-driven classification of 46 medicinal macrofungi species: comparative analysis, fusion strategies, and explainability insights.
AU - Mustafa Sevindik
AU - Aras Fahrettin Korkmaz
AU - Fatih Ekinci
AU - Eda Kumru
AU - Tetiana Krupodorova
AU - Ömer Burak Altindal
AU - Mehmet Serdar Güzel
AU - Ilgaz Akata
PY - 2026
DA - 2026/08/05
JO - BMC bioinformatics
DO - 10.1186/s12859-026-06584-z
UR - https://europepmc.org/article/MED/42967984
AB - Accurate identification of medicinal and wild macrofungi remains a major challenge due to high intra-species variability, visually overlapping morphologies, and inconsistent field-level imaging conditions. Existing deep learning studies typically focus on limited taxonomic scopes (5-12 classes) and lack systematic comparisons across modern convolutional and transformer architectures, leaving a significant gap in fine-grained fungal classification. This study introduces a comprehensive evaluation framework covering 46 macrofungal species, integrating 18 single backbones and three feature-level fusion strategies under a unified, fully reproducible protocol. All models were trained with identical preprocessing, stratified partitions, deterministic seeds, and a two-stage fine-tuning pipeline to ensure fair architectural comparison. Among single models, ConvNeXt-Base achieved the strongest performance with 96.59% accuracy, 0.966 macro-F1, 0.965 MCC, 0.9997 macro-AUC, and 0.1054 log-loss, exhibiting a near-diagonal confusion matrix where 25 classes reached perfect recall. Although feature-level fusion did not exceed the best single model in overall accuracy, it improved stability for several visually difficult taxa. The best ensemble, ConvNeXt-Base + DeiT-Base, achieved 95.85% accuracy, 0.958 macro-F1, and 0.9995 macro-AUC, remaining close to the ConvNeXt-Base baseline while providing complementary class-level robustness. The reported latency of ≈0.0021 s/image refers only to the precomputed-feature mode, in which backbone features are already cached and only the shared MLP head is evaluated. Therefore, this value should not be interpreted as full end-to-end inference latency for newly acquired field images. Explainability analyses using Grad-CAM and Integrated Gradients demonstrated that both single and fused models rely on biologically meaningful structures such as gill edges, pore surfaces, cap margins, and hymenophore textures. Overall, this study provides one of the most extensive quantitative benchmarks for macrofungal recognition to date and establishes a reproducible, architecture-agnostic foundation for future multimodal, large-scale, and explainable AI systems in fungal taxonomy.
SN - 1471-2105
ER -