Classification of Crystallization Images of Pharmaceutical Raw Materials Using Convolutional Neural Network Algorithm

Anton   Yudhana; Julia Mega  Reski

doi:10.59395/ijadis.v6i3.1440

Authors

Anton Yudhana Universitas Ahmad Dahlan
Julia Mega Reski Universitas Ahmad Dahlan

DOI:

https://doi.org/10.59395/ijadis.v6i3.1440

Keywords:

CNN, crystallization,, deep learning, pharmaceutical raw materials, Image Classification

Abstract

The rapid advancement of artificial intelligence (AI) has opened new opportunities for automation in the pharmaceutical industry, particularly in the classification of raw drug materials. Manual classification methods are time-consuming and prone to human error, highlighting the need for reliable automated solutions. This study applied a deep learning approach for classifying crystallization images of pharmaceutical raw materials using a Convolutional Neural Network (CNN). A dataset of 300 crystallization images of Nicotinamide and Ferulic Acid was obtained through hot-stage microscopy, preprocessed with normalization, resizing, and augmentation, and divided into training, validation, and testing subsets. The CNN model was trained for 10 epochs and evaluated using a confusion matrix and standard performance metrics (accuracy, precision, recall, and F1-score). The model achieved perfect recall for Ferulic Acid and 90% recall with 100% precision for Nicotinamide, resulting in an overall accuracy of 95%. While these results are promising, the relatively small dataset may limit generalization, and further validation with larger or external datasets is required. The findings indicate that CNN-based methods hold strong potential for automating crystallization classification, improving pharmaceutical quality control, and reducing reliance on manual assessment, in line with recent advances in medical and pharmaceutical image analysis.

1215 706

Downloads

Download data is not yet available.

References

[1] F. Artusio, R. Contreras-Montoya, and J. A. Gavira, Advances in Pharmaceutical Crystals: Control over Nucleation and Polymorphism, Crystals, vol. 14, no. 9, pp. 46, 2024, doi: 10.3390/cryst14090805. DOI: https://doi.org/10.3390/cryst14090805

[2] A. Cote et al., Perspectives on the Current State, Challenges, and Opportunities in Pharmaceutical Crystallization Process Development, Cryst. Growth Des., vol. 20, no. 12, pp. 75687581, 2020, doi: 10.1021/acs.cgd.0c00847. DOI: https://doi.org/10.1021/acs.cgd.0c00847

[3] I. Indra, R. Rahman, and R. Yulianti, Karakterisasi Karbamazepin Hasil Rekristalisasi Berbagai Pelarut Organik dengan Metode Slow Evaporation, J. Farm. Dan Ilmu Kefarmasian Indones., vol. 8, no. 3, p. 227, 2021, doi: 10.20473/jfiki.v8i32021.227-234. DOI: https://doi.org/10.20473/jfiki.v8i32021.227-234

[4] M. Lambert et al., Modular microfluidic platform for solubility measurement, nucleation statistics and polymorph screening of active pharmaceutical ingredients: Irbesartan, Rimonabant, Aripiprazole and Sulfathiazole, J. Cryst. Growth, vol. 616, 2023, doi: 10.1016/j.jcrysgro.2023.127252. DOI: https://doi.org/10.1016/j.jcrysgro.2023.127252

[5] T. Jiko, Crystallization in Pharmaceutical Manufacturing: Optimizing Drug Stability and Bioavailability, pp. 12, doi: 10.35248/2157-7064.25.16.602.Citation.

[6] A. Kumar, P. Singh, and A. Nanda, Hot stage microscopy and its applications in pharmaceutical characterization, Appl. Microsc., vol. 50, no. 1, 2020, doi: 10.1186/s42649-020-00032-9. DOI: https://doi.org/10.1186/s42649-020-00032-9

[7] M. Mousavi and S. Hosseini, A deep convolutional neural network approach using medical image classification, BMC Med. Inform. Decis. Mak., vol. 24, no. 1, 2024, doi: 10.1186/s12911-024-02646-5. DOI: https://doi.org/10.1186/s12911-024-02646-5

[8] O. N. F. King, K. E. Levik, J. Sandy, and M. Basham, CHiMP: deep-learning tools trained on protein crystallization micrographs to enable automation of experiments, Acta Crystallogr. Sect. D Struct. Biol., vol. 80, no. Pt 10, pp. 744764, 2024, doi: 10.1107/S2059798324009276. DOI: https://doi.org/10.1107/S2059798324009276

[9] A. Muis, S. Sunardi, and A. Yudhana, Medical image classification of brain tumor using convolutional neural network algorithm, J. Infotel, vol. 15, no. 3, pp. 227232, 2023, doi: 10.20895/infotel.v15i3.964. DOI: https://doi.org/10.20895/infotel.v15i3.964

[10] H. Lai, K. Sevrani, and S. Iqbal, Deep learning approaches for classification tasks in medical X-ray, MRI, and ultrasound images: a scoping review, BMC Med. Imaging, vol. 25, no. 1, 2025, doi: 10.1186/s12880-025-01701-5. DOI: https://doi.org/10.1186/s12880-025-01701-5

[11] A. Yilmaz, A. A. Demircali, S. Kocaman, and H. Uvet, Comparison of Deep Learning and Traditional Machine Learning Techniques for Classification of Pap Smear Images, 2020, [Online]. Available: http://arxiv.org/abs/2009.06366

[12] P. Wang, E. Fan, and P. Wang, Comparative analysis of image classification algorithms based on traditional machine learning and deep learning, Pattern Recognit. Lett., vol. 141, pp. 6167, 2021, doi: 10.1016/j.patrec.2020.07.042. DOI: https://doi.org/10.1016/j.patrec.2020.07.042

[13] Y. Kang et al., An Enhanced Deep Learning-Based Pharmaceutical Crystal Detection with Regional Filtering, Crystals, vol. 14, no. 8, 2024, doi: 10.3390/cryst14080709. DOI: https://doi.org/10.3390/cryst14080709

[14] J. Milne, C. Qian, D. Hargreaves, Y. Wang, and J. Wilson, Not getting in too deep: A practical deep learning approach to routine crystallisation image classification, PLoS One, vol. 18, no. 3 March, pp. 116, 2023, doi: 10.1371/journal.pone.0282562. DOI: https://doi.org/10.1371/journal.pone.0282562

[15] F. E. Pujiono, J. Ekowati, T. Amrillah, and D. Setyawan, Ferulic Acid-Nicotinamide Cocrystal: Synthesis, Experimental, and Computation Study, Sci. Technol. Indones., vol. 10, no. 2, pp. 402410, 2025, doi: 10.26554/sti.2025.10.2.402-410. DOI: https://doi.org/10.26554/sti.2025.10.2.402-410

[16] D. Characteristics, Tropical Journal of Natural Product Research Enhancing the Pharmaceutical Properties of Ibuprofen through Spherical Agglomeration-Co-Crystallization with Nicotinamide: A Comprehensive Study of, vol. 8, no. July, pp. 77157721, 2024. DOI: https://doi.org/10.26538/tjnpr/v8i8.48

[17] R. S. Jebur, M. H. B. M. Zabil, D. A. Hammood, and L. K. Cheng, A comprehensive review of image denoising in deep learning, Multimed. Tools Appl., vol. 83, no. 20, pp. 5818158199, 2024, doi: 10.1007/s11042-023-17468-2. DOI: https://doi.org/10.1007/s11042-023-17468-2

[18] I. Corley, C. Robinson, R. Dodhia, J. M. Lavista Ferres, and P. Najafirad, Revisiting pre-trained remote sensing model benchmarks: Resizing and normalization matters, IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. Work., pp. 31623172, 2024, doi: 10.1109/CVPRW63382.2024.00322. DOI: https://doi.org/10.1109/CVPRW63382.2024.00322

[19] E. Jain, J. Neeraja, B. Banerjee, and P. Ghosh, A Diagnostic Approach to Assess the Quality of Data Splitting in Machine Learning, 2022, [Online]. Available: http://arxiv.org/abs/2206.11721

[20] H. Bichri, A. Chergui, and M. Hain, Investigating the Impact of Train / Test Split Ratio on the Performance of Pre-Trained Models with Custom Datasets, Int. J. Adv. Comput. Sci. Appl., vol. 15, no. 2, pp. 331339, 2024, doi: 10.14569/IJACSA.2024.0150235. DOI: https://doi.org/10.14569/IJACSA.2024.0150235

[21] I. E. Tampu, A. Eklund, and N. Haj-Hosseini, Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images, Sci. Data, vol. 9, no. 1, pp. 19, 2022, doi: 10.1038/s41597-022-01618-6. DOI: https://doi.org/10.1038/s41597-022-01618-6

[22] S. Chatterjee and P. Zielinski, On the Generalization Mystery in Deep Learning, 2022, [Online]. Available: http://arxiv.org/abs/2203.10036

[23] S. Jelassi and Y. Li, Towards understanding how momentum improves generalization in deep learning, Proc. Mach. Learn. Res., vol. 162, pp. 996510040, 2022.

[24] M. Lashkari and A. Gheibi, Lipschitzness effect of a loss function on generalization performance of deep neural networks trained by Adam and AdamW optimizers, AUT J. Math. Comput., vol. 5, no. 4, pp. 361375, 2024, doi: 10.22060/ajmc.2023.22182.1139.

[25] L. Zhu, C. Liu, A. Radhakrishnan, and M. Belkin, Catapults in SGD: spikes in the training loss and their impact on generalization through feature learning, Proc. Mach. Learn. Res., vol. 235, pp. 6247662509, 2024.

[26] M. Wang and C. Ma, Generalization Error Bounds for Deep Neural Networks Trained by SGD, 2022, [Online]. Available: http://arxiv.org/abs/2206.03299

[27] L. Zewen, L. Fan, Y. Wenjie, P. Shouheng, and Z. Jun, A survey of convolutional neural networks: analysis, applications, and prospects, IEEE Trans. neural networks Learn. Syst., vol. 33, no. 12, pp. 69997019, 2021. DOI: https://doi.org/10.1109/TNNLS.2021.3084827

[28] Y. L. Ma and M. Rho, Towards the hadronquark continuity via a topology change in compact stars, Prog. Part. Nucl. Phys., vol. 113, pp. 187, 2020, doi: 10.1016/j.ppnp.2020.103791. DOI: https://doi.org/10.1016/j.ppnp.2020.103791

[29] S. Minaee, Y. Boykov, F. Porikli, A. Plaza, N. Kehtarnavaz, and D. Terzopoulos, Image Segmentation Using Deep Learning: A Survey, IEEE Trans. Pattern Anal. Mach. Intell., vol. 44, no. 7, pp. 35233542, 2022, doi: 10.1109/TPAMI.2021.3059968. DOI: https://doi.org/10.1109/TPAMI.2021.3059968

[30] J. Grtler et al., Neo: Generalizing Confusion Matrix Visualization to Hierarchical and Multi-Output Labels, vol. 1, no. 1. Association for Computing Machinery, 2022. doi: 10.1145/3491102.3501823. DOI: https://doi.org/10.1145/3491102.3501823

[31] K. Riehl, M. Neunteufel, and M. Hemberg, Hierarchical confusion matrix for classification performance evaluation, J. R. Stat. Soc. Ser. C Appl. Stat., vol. 72, no. 5, pp. 13941412, 2023, doi: 10.1093/jrsssc/qlad057. DOI: https://doi.org/10.1093/jrsssc/qlad057

[32] A. Farhan AlShammari, Implementation of Model Evaluation using Confusion Matrix in Python, Int. J. Comput. Appl., vol. 186, no. 50, pp. 9758887, 2024. DOI: https://doi.org/10.5120/ijca2024924236

[33] S. Kanakala and S. Ningappa, Detection and Classification of Diseases in Multi-Crop Leaves using LSTM and CNN Models, J. Innov. Image Process., vol. 7, no. 1, pp. 161181, 2025, doi: 10.36548/jiip.2025.1.008. DOI: https://doi.org/10.36548/jiip.2025.1.008

[34] D. Chicco and G. Jurman, The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation, BMC Genomics, vol. 21, no. 1, pp. 113, 2020, doi: 10.1186/s12864-019-6413-7. DOI: https://doi.org/10.1186/s12864-019-6413-7

[35] A. Tharwat, Classification assessment methods, Appl. Comput. Informatics, vol. 17, no. 1, pp. 168192, 2018, doi: 10.1016/j.aci.2018.08.003. DOI: https://doi.org/10.1016/j.aci.2018.08.003

[36] H. Kaur and D. Kaur Sandhu, Evaluating the Effectiveness of the Proposed System Using F1 Score, Recall, Accuracy, Precision and Loss Metrics Compared to Prior Techniques, Int. J. Commun. Networks Inf. Secur., vol. 15, no. 4, pp. 368383, 2023, [Online]. Available: https://ijcnis.org

[37] M. Sitarz, Extending F1 Metric, Probabilistic Approach, Adv. Artif. Intell. Mach. Learn., vol. 3, no. 2, pp. 10251038, 2023, doi: 10.54364/AAIML.2023.1161. DOI: https://doi.org/10.54364/AAIML.2023.1161

[38] L. Alzubaidi et al., Review of deep learning: concepts, CNN architectures, challenges, applications, future directions, vol. 8, no. 1. Springer International Publishing, 2021. doi: 10.1186/s40537-021-00444-8. DOI: https://doi.org/10.1186/s40537-021-00444-8

[39] F. A. Mohammed, K. K. Tune, B. G. Assefa, M. Jett, and S. Muhie, Medical Image Classifications Using Convolutional Neural Networks: A Survey of Current Methods and Statistical Modeling of the Literature, Mach. Learn. Knowl. Extr., vol. 6, no. 1, pp. 699735, 2024, doi: 10.3390/make6010033. DOI: https://doi.org/10.3390/make6010033

[40] A. Teramoto, A. Michiba, Y. Kiriyama, E. Sakurai, R. Shiroki, and T. Tsukamoto, Automated Classification of Urinary Cells: Using Convolutional Neural Network Pre-trained on Lung Cells, Appl. Sci., vol. 13, no. 3, 2023, doi: 10.3390/app13031763. DOI: https://doi.org/10.3390/app13031763