Evaluation of a hybrid AI model for the automatic identification of pollen morphological features for apitherapy applications

Yıl 2025, Cilt: 8 Sayı: 2, 268 - 294, 31.12.2025

Uğur Şevik

https://doi.org/10.35206/jan.1826463

Öz

Melissopalynology is the gold standard for authenticating honey but traditional microscopic analysis is time-consuming and subjective. This study evaluates a hybrid artificial intelligence approach to automate pollen classification using the comprehensive POLLEN73S dataset, which features 73 distinct pollen types from the Brazilian Savanna. To address class imbalance, the dataset was expanded to 7300 images using data augmentation. We extracted morphological features using three pre-trained deep learning models (ResNet50, EfficientNetB0, MobileNetV2) and classified them using 17 traditional machine learning algorithms. The hybrid model combining ResNet50 features with Linear Discriminant Analysis (LDA) achieved the highest accuracy of 97.00%. Error analysis indicated that misclassifications were concentrated among taxonomically similar genera, such as Serjania, due to shared exine structures. These results demonstrate that the proposed hybrid model offers a highly accurate and scalable solution for laboratory-based honey authentication, provided it is integrated with debris detection systems to handle real-world samples.

Anahtar Kelimeler

Melissopalynology , Pollen Morphology , POLLEN73S , Honey Authentication , Hybrid Classification , ResNet50 , Linear Discriminant Analysis

Apiterapi uygulamaları için polen morfolojik özelliklerinin otomatik tanımlanmasına yönelik hibrit bir yapay zeka modelinin değerlendirilmesi

Öz

Melissopalinoloji, balın kimliğini doğrulamada altın standarttır; ancak geleneksel mikroskobik analiz zaman alıcı ve öznel bir süreçtir. Bu çalışma, Brezilya Savanası'na ait 73 farklı polen tipini içeren kapsamlı POLLEN73S veri setini kullanarak, polen sınıflandırmasını otomatikleştirmeye yönelik hibrit bir yapay zeka yaklaşımını değerlendirmektedir. Sınıf dengesizliğini gidermek amacıyla, veri artırma teknikleri kullanılarak veri seti 7300 görüntüye genişletilmiştir. Çalışmada, üç farklı önceden eğitilmiş derin öğrenme modeli (ResNet50, EfficientNetB0, MobileNetV2) kullanılarak morfolojik öznitelikler çıkarılmış ve bu öznitelikler 17 geleneksel makine öğrenmesi algoritması ile sınıflandırılmıştır. ResNet50 özniteliklerini Doğrusal Diskriminant Analizi (LDA) ile birleştiren hibrit model, %97,00 ile en yüksek doğruluk oranına ulaşmıştır. Hata analizi, hatalı sınıflandırmaların, paylaşılan ekzin yapıları nedeniyle Serjania gibi taksonomik olarak benzer cinsler arasında yoğunlaştığını göstermiştir. Bu sonuçlar, önerilen hibrit modelin; gerçek dünya örneklerini işleyebilmek adına kalıntı tespit sistemleriyle entegre edilmesi koşuluyla, laboratuvar tabanlı bal kimlik doğrulaması için yüksek doğruluklu ve ölçeklenebilir bir çözüm sunduğunu ortaya koymaktadır.

Anahtar Kelimeler

Melissopalinoloji , Polen Morfolojisi , POLLEN73S , Bal Kimlik Doğrulaması , Hibrit Sınıflandırma , ResNet50 , Doğrusal Diskriminant Analizi

Kaynakça

• Adaïmé, M. É., Kong, S., & Punyasena, S. W. (2024). Deep learning approaches to the phylogenetic placement of extinct pollen morphotypes. PNAS Nexus, 3(1), pgad419.
• Alissandrakis, E., Tsiknakis, N., Savvidaki, E., Kafetzopoulos, S., Manikis, G., Vidakis, N., & Marias, K. (2021). Cretan Pollen Dataset v1 (CPD-1). Zenodo. https://doi.org/10.5281/zenodo.4756360
• Astolfi, G., & Gonçalves, A. B. (2020). POLLEN73S. Figshare. https://doi.org/10.6084/m9.figshare.12536573.v1
• Astolfi, G., Gonçalves, A. B., Menezes, G. V., Borges, F. S. B., Astolfi, A. C. M. N., Matsubara, E. T.,... & Pistori, H. (2020). POLLEN73S: An image dataset for pollen grains classification. Ecological Informatics, 60, 101165. https://doi.org/10.1016/j.ecoinf.2020.101165
• Battiato, S., Guarnera, F., Ortis, A., Trenta, F., Ascari, L., Siniscalco, C.,... & Suárez, E. (2020). Pollen Grain Classification Challenge 2020 Challenge Report. In A. Del Bimbo et al. (Eds.), ICPR 2020 Workshops, LNCS 12668 (ss. 469-479). Springer. https://doi.org/10.1007/978-3-030-68793-9_34
• Bicudo de Almeida-Muradian, L., Stramm, K. M., & Estevinho, L. M. (2020). Melissopalynology. In Honey Analysis (ss. 1-28). Springer.
• Boldeanu, M. (2022). Automatic pollen classification using deep learning techniques (Ph.D. Thesis Summary). Politehnica University of Bucharest. Cascante-Bonilla, P., Tan, F., Qi, Y., & Li, V. (2020). Curriculum labeling: A novel approach to semi-supervised learning. arXiv preprint arXiv:2001.06001.
• Chippa, P., Hu, S., Pound, M., Yawar, S. A., & Baniulis, D. (2025). Honey authentication using AI-based pollen analysis: a UK review. British Food Journal. (Basımda).
• Chollet, F. (2017). Xception: Deep learning with depthwise separable convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) (ss. 1251-1258).
• Daood, A., Dulam, C. S., & Haci, H. (2016). Pollen grain classification using deep learning. IEEE Conference on Advances in Electrical, Electronic and Systems Engineering (ICAEES).
• Erdtman, G. (1960). The acetolysis method: A revised description. Svensk Botanisk Tidskrift, 54, 561-564.
• Erdtman, G. (1966). Pollen morphology and plant taxonomy: Angiosperms. Hafner Publishing Company. France, I., Duller, A. W. G., Duller, G. A. T., & Lamb, H. F. (2000). A new approach to automated pollen analysis. Quaternary Science Reviews, 19(6), 537-546.
• Gallardo-Caballero, R., Valiente-González, J. M., & González-Alonso, V. (2019). Detection of pollen grains in digital images using a convolutional neural network. Pattern Recognition Letters, 125, 223-230. Gallardo, M., Valiente, J. M., Gonzalez-Alonso, V., & Casanas, M. (2024). CAPI Pollen DB2 dataset. Data in Brief, 52, 109961.
• García, M. E., Mora, M. R., & Barboza, C. G. (2012). Pollen grain classification using HMM. IEEE Signal Processing Society International Conference on Acoustics, Speech and Signal Processing (ICASSP).
• Garga, B., Abboubakar, H., Sourpele, R. S., Gwet, D. L. L., & Bitjoka, L. (2024). Pollen Grain Classification Using Some Convolutional Neural Network Architectures. Journal of Imaging, 10(7), 158. https://doi.org/10.3390/jimaging10070158
• Gimenez, B., Joannin, S., Pasquet, J., Beaufort, L., Gally, Y., de Garidel-Thoron, T.,... & Peyron, O. (2024). A user-friendly method to get automated pollen analysis from environmental samples. New Phytologist, 243(2), 797-810. https://doi.org/10.1111/nph.19857
• Gonçalves, A. B., Souza, J. S., Silva, G. G. D., Cereda, M. P., Pott, A., Naka, M. H., & Pistori, H. (2016). Feature Extraction and Machine Learning for the Classification of Brazilian Savannah Pollen Grains. PloS one, 11(6), e0157044. https://doi.org/10.1371/journal.pone.0157044
• Halbritter, H., Ulrich, S., Grímsson, F., Weber, M., Zetter, R., Waanders, M.,... & Svojtka, M. (2018). Illustrated pollen terminology. Springer.
• He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) (ss. 770-778).
• Hesse, M., Halbritter, H., Zetter, R., Weber, M., Buchner, R., Frosch-Radivo, A., & Ulrich, S. (2009). Pollen terminology: An illustrated handbook. Springer. Holt, K. A. (2020). Classifynder 46: A dataset for automated pollen classification. Zenodo.
• Hu, J., Shen, L., & Sun, G. (2018). Squeeze-and-excitation networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) (ss. 7132-7141).
• Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017). Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) (ss. 4700-4708).
• Ioffe, S., & Szegedy, C. (2015). Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167.
• Karras, T., Laine, S., Aittala, M., Hellsten, J., Lehtinen, J., & Aila, T. (2020). Analyzing and improving the image quality of StyleGAN. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (CVPR) (ss. 8110-8119).
• Khalane, J. S., Gawande, N. D., Raman, S., & Sankaranarayanan, S. (2025). IMPORTANT: Advanced Pollen Classification of Indian Medicinal Plants through SEM and Computer Vision. bioRxiv. https://doi.org/10.1101/2025.01.08.631879
• Kong, S., Punyasena, S. W., & Fowlkes, C. C. (2016). Spatially aware sparse coding for fossil pollen identification. Neural Information Processing Systems (NIPS).
• Kubera, Y., Samek, W., & Stacewicz, P. (2021). Pollen grain detection using YOLOv5. IEEE International Conference on Image Processing (ICIP).
• LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444.
• Li, C., Polling, M., Cao, L., Gravendeel, B., & Verbeek, F. J. (2023). Analysis of automatic image classification methods for Urticaceae pollen classification. Neurocomputing, 522, 181-193. https://doi.org/10.1016/j.neucom.2022.11.042
• Long, Y., Sun, W., Sun, N., Wang, W., Li, C., & Yin, S. (2025). HieraEdgeNet: A multi-scale edge-enhanced framework for automated pollen recognition. arXiv preprint arXiv:2506.07637.
• Lu, L., Jiao, B. H., Qin, F., Xie, G., Lu, K. Q., Li, J. F.,... & Wang, Y. F. (2022). Artemisia pollen dataset for exploring the potential ecological indicators in deep time. Earth System Science Data, 14(9), 3961-3995. https://doi.org/10.5194/essd-14-3961-2022
• Mahbod, A., Schaefer, G., Ecker, R., & Ellinger, I. (2021). Classification of pollen grains using deep convolutional neural networks. IEEE International Conference on Image Processing (ICIP).
• Mahmood, T., Choi, J., & Park, K. R. (2023). Artificial intelligence-based classification of pollen grains using attention-guided pollen features aggregation network. Journal of King Saud University-Computer and Information Sciences, 35(1), 740-756. https://doi.org/10.1016/j.jksuci.2023.01.013
• Manikis, G., Kafetzopoulos, S., Tsiknakis, N., & Marias, K. (2019). Pollen grain classification using geometrical and textural features. IEEE 19th International Conference on Bioinformatics and Bioengineering (BIBE).
• Marques, D., Via do Pico, G. M., Nakajima, J. N., & Dematteis, M. (2021). Pollen morphology and its systematic value to southern South American species of Lepidaploa (Vernonieae: Asteraceae). Rodriguésia, 72, e01412019. https://doi.org/10.1590/2175-7860202172017
• Mills, B., Zervas, M. N., & Grant-Jacob, J. A. (2025). Pollen image manipulation and projection using latent space. Frontiers in Plant Science, 16, 1539128. https://doi.org/10.3389/fpls.2025.1539128
• Nair, V., & Hinton, G. E. (2010). Rectified linear units improve restricted boltzmann machines. ICML.
• Olsson, O., Kjell, A., & Kjell, H. (2021). Automated pollen analysis using deep learning. Grana, 60(1), 1-15.
• Pozo-Banos, M. D., et al. (2012). Pollen grain classification using a feature selection method. Applied Mathematics & Information Sciences, 6(1), 163-172.
• Pound, M. J., Riding, J. B., & Donders, T. H. (2023). Melissopalynology. In Applying palynology (ss. 45-66). CRC Press. Punt, W., Hoen, P. P., Blackmore, S., Nilsson, S., & Le Thomas, A. (2007). Glossary of pollen and spore terminology. Review of Palaeobotany and Palynology, 143(1-2), 1-81.
• Punyasena, S. W., Tcheng, D., Wesseln, C., & Mueller, P. (2012). Classifying airborne pollen, fungal spores, and other aerosolized biological particles: A machine learning approach. Atmospheric Environment, 59, 259-266.
• Riding, J. B. (2021). A guide to palynological preparation techniques. Palynology, 45(sup1), 1-62.
• Romero, M., Tcheng, D., & Punyasena, S. W. (2020). Automated classification of legume pollen using optical superresolution microscopy and deep learning. Paleobiology, 46(4), 527-541.
• Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S.,... & Fei-Fei, L. (2015). ImageNet large scale visual recognition challenge. International Journal of Computer Vision, 115(3), 211-252.
• Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L. C. (2018). MobileNetV2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) (ss. 4510-4520).
• Sevillano, V., & Aznarte, J. L. (2018). Improving classification of pollen grain images of the POLEN23E dataset through three different applications of deep learning convolutional neural networks. PloS one, 13(9), e0201807. https://doi.org/10.1371/journal.pone.0201807
• Sevillano, V., Aznarte, J. L., & Fernández, C. (2020). Classification of 46 different pollen grain types using deep learning. IEEE Access, 8, 176140-176152.
• Simonyan, K., & Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
• Soares, J. C. D. S., Aires, K. R. T., Bendini, J. D. N., Brandão, W. V. B., & Veras, R. D. M. S. (2025). Enhanced Pollen Classification via Hybrid Neural Network With Attention Mechanism and View Separation: An Equatorial and Polar Approach. IEEE Access, 13, 77365-77373. https://doi.org/10.1109/ACCESS.2025.3562316
• Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016). Rethinking the inception architecture for computer vision. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) (ss. 2818-2826).
• Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017). Inception-v4, inception-resnet and the impact of residual connections on learning. In AAAI Conference on Artificial Intelligence (Vol. 31, No. 1).
• Tan, M., & Le, Q. V. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. ICML.
• Ticay-Rivas, J. R., Rojas-Alvarado, B., & García-Mora, M. E. (2011). Classification of pollen grains using multilayer neural networks. IEEE International Conference on Electronics, Communications and Computing (CONIELECOMP). Treloar, W. J., Taylor, P. E., & Flenley, J. R. (2004). Pollen recognition from the automated analysis of texture. Review of Palaeobotany and Palynology, 131(1-2), 1-11. Tsiknakis, N., Savvidaki, E., Kafetzopoulos, S., Manikis, G., Vidakis, N., Marias, K., & Alissandrakis, E. (2021). Segmenting 20 Types of Pollen Grains for the Cretan Pollen Dataset v1 (CPD-1). Applied Sciences, 11(14), 6657. https://doi.org/10.3390/app11146657
• Tsiknakis, N., Savvidaki, E., Manikis, G. C., Gotsiou, P., Remoundou, I., Marias, K.,... & Vidakis, N. (2022). Pollen Grain Classification Based on Ensemble Transfer Learning on the Cretan Pollen Dataset. Plants, 11(7), 919. https://doi.org/10.3390/plants11070919
• Von Der Ohe, W., Persano Oddo, L., Piana, M. L., Morlot, M., & Martin, P. (2004). Harmonized methods of melissopalynology. Apidologie, 35(Suppl 1), S18-S25.
• Woo, S., Park, J., Lee, J. Y., & Kweon, I. S. (2018). CBAM: Convolutional block attention module. In Proceedings of the European conference on computer vision (ECCV) (ss. 3-19).
• Zhang, T., & Mao, L. (2026). Deep learning of pollen images under low annotation costs: joint optimization of morphological features and training and prediction strategies. Review of Palaeobotany and Palynology, 344, 105458. https://doi.org/10.1016/j.revpalbo.2025.105458
• Zolfaghari, M., & Sajedi, H. (2024). Automated classification of pollen grains microscopic images using cognitive attention based on human Two Visual Streams Hypothesis. PloS one, 19(11), e0309674. https://doi.org/10.1371/journal.pone.0309674
• Zoph, B., & Le, Q. V. (2017). Neural architecture search with reinforcement learning. arXiv preprint arXiv:1611.01578. Zoph, B., Vasudevan, V., Shlens, J., & Le, Q. V. (2018). Learning transferable architectures for scalable image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) (ss. 8697-8710).