Hybrid Deep Feature Extraction and Machine Learning for Potato Disease Classification
DOI:
https://doi.org/10.24252/shift.v6i2.285Keywords:
EfficientNetB3, SVM, XGBoost, Random ForestAbstract
Accurate classification of potato leaf diseases is crucial for improving agricultural productivity and reducing crop losses. This study proposes a hybrid classification framework that combines EfficientNetB3 as a deep feature extractor with classical machine learning classifiers, namely Support Vector Machine (SVM), Random Forest (RF), and XGBoost, to overcome the limitations of end-to-end convolutional neural networks in terms of computational cost and training time. Deep features extracted by EfficientNetB3 are used as inputs for the respective classifiers, and the proposed models are evaluated using accuracy, precision, recall, F1-score, and training time. Experimental results demonstrate that the hybrid models achieve superior performance and efficiency compared to the baseline EfficientNetB3 with Softmax classifier. Among all methods, EfficientNetB3 + XGBoost attains the highest accuracy of 0.966 with reduced training time, while EfficientNetB3 + SVM offers competitive accuracy with the fastest training process. These findings indicate that integrating deep feature extraction with traditional machine learning classifiers provides an effective and computationally efficient approach for potato leaf disease classification, making it suitable for practical implementation in precision agriculture.
Downloads
References
[1] V. Sharma, A. Kaur, and A. Kumar, “Early blight disease management of potato,” Potato Res., vol. 68, no. 2, pp. 1869–1892, 2025, doi: 10.1007/s11540-024-09816-y.
[2] F. Marisa, M. A. R. Ramadhani, A. L. Maukar, P. Savitri, Saepudin, and A. Syafariah, “Physio-logical symptom-based classification of potato leaf diseases using K-nearest neighbor,” in Proc. 2025 19th Int. Conf. Telecommun. Syst., Serv., Appl. (TSSA), 2025, pp. 1–6, doi: 10.1109/TSSA68467.2025.11303840.
[3] F. M. Farag and A. Khalifa, “Biocontrol potential of endophytic Bacillus spp. against early blight disease in potato (Solanum tuberosum),” Biol. Control, vol. 210, p. 105897, Nov. 2025, doi: 10.1016/j.biocontrol.2025.105897.
[4] J. Li et al., “Transcriptome analysis of potato (Solanum tuberosum L.) seedlings with vary-ing resistance levels reveals diverse molecular pathways in early blight resistance,” Plants, vol. 14, no. 15, p. 2422, 2025, doi: 10.3390/plants14152422.
[5] A. Kumar and V. K. Patel, “Classification and identification of disease in potato leaf using hierarchical based deep learning convolutional neural network,” Multimed. Tools Appl., vol. 82, no. 20, pp. 31101–31127, 2023, doi: 10.1007/s11042-023-14663-z.
[6] A. T. Mulugeta, W. Jifara, E. Bogale, T. Desiyo, and A. Mokonnen, “Early detection and clas-sification of potato leaf disease using convolutional neural networks,” Appl. Comput. In-tell. Soft Comput., vol. 2025, no. 1, p. 7614841, 2025, doi: 10.1155/acis/7614841.
[7] M. Shendage, K. Golekar, A. Jambhale, and A. Bang, “Potato disease classification using deep learning,” in Proc. 2025 Int. Conf. Mach. Learn. Auton. Syst. (ICMLAS), 2025, pp. 342–350, doi: 10.1109/ICMLAS64557.2025.10967792.
[8] K. Kayathri, P. Sasikala, S. B. Rajakumari, K. S. Priya, M. R. Ramya, and S. Mohanambal, “Enhanced tomato leaf disease classification using the EfficientNetB3 and ResNet50 fu-sion,” in Proc. 2025 14th Int. Conf. Syst. Model. Adv. Res. Trends (SMART), 2025, pp. 219–226, doi: 10.1109/SMART66937.2025.11389623.
[9] L. K. Kumari et al., “A deep learning based approach for tomato leaf disease detection us-ing EfficientNetB3,” in Proc. 2025 Int. Conf. Cogn., Green Ubiquitous Comput. (IC-CGU), 2025, pp. 1–6, doi: 10.1109/IC-CGU67042.2025.11338013.
[10] G. R. Panigrahi, G. S. Shiny, G. V. Londhe, S. Sharma, A. Misra, and D. Bhardwaj, “Automat-ed tomato leaf disease detection using EfficientNetB3: A deep learning approach for agri-cultural diagnostics,” in Proc. 2025 Int. Conf. Decis. Aid Sci. Appl. (DASA), 2025, pp. 1580–1585, doi: 10.1109/DASA68193.2025.11498966.
[11] M. A. Patil, M. Manohar, C. Laxuman, K. Parane, B. M. Dodamani, and G. Sunkad, “Hybrid deep learning-based potato and tomato leaf disease classification,” in Advances in Data-Driven Computing and Intelligent Systems (Lecture Notes in Networks and Systems). Sin-gapore: Springer, 2024, pp. 157–174, doi: 10.1007/978-981-99-9524-0_13.
[12] T. Illakiya, A. V. Joel, N. D. Nahata, S. Sukumar, and S. P. Chelladurai, “A hybrid deep learn-ing model for potato leaf disease classification using transfer learning and attention mech-anism,” in Proc. 2025 8th Int. Conf. Comput. Methodol. Commun. (ICCMC), 2025, pp. 837–842, doi: 10.1109/ICCMC65190.2025.11140741.
[13] L. Priyadharshini, P. Kalaiselvi, S. Sujatha, S. Uma Maheswari, V. A. Kandaswamy, and R. Suganthi, “HNCS: A robust development of deep learning based potato plant disease iden-tification using hybrid neural classification strategy,” in Proc. 2024 4th Asian Conf. Innov. Technol. (ASIANCON), 2024, pp. 1–5, doi: 10.1109/ASIANCON62057.2024.10838204.
[14] A. R. Nalla, D. K. Basha, A. Archana, and K. S. Rao, “Plant disease classification using Effi-cientNetB3 for tomato leaf,” in Proc. 2025 IEEE 1st Int. Conf. Recent Trends Comput. Smart Mobility (RCSM), 2025, pp. 1–6, doi: 10.1109/RCSM67767.2025.11507776.
[15] H. I. Peyal and M. N. I. Mondal, “Potato leaf disease classification using lightweight CNN-SVM and real-world web-based tools,” in Proc. 2025 Int. Conf. Elect., Comput. Commun. Eng. (ECCE), 2025, pp. 1–5, doi: 10.1109/ECCE64574.2025.11013900.
[16] S. Chauhan and K. S. Kumar, “AI-driven potato disease classification: ResNet50 for detect-ing late blight in crop leaves,” in Proc. 2025 3rd Int. Conf. Augment. Intell. Sustain. Syst. (ICAISS), 2025, pp. 793–798, doi: 10.1109/ICAISS61471.2025.11042051.
[17] A. Singh and H. Kaur, “Potato plant leaves disease detection and classification using ma-chine learning methodologies,” IOP Conf. Ser.: Mater. Sci. Eng., vol. 1022, no. 1, p. 012121, 2021, doi: 10.1088/1757-899X/1022/1/012121.
[18] A. H. T. Rusli, B. C. C. Meng, N. S. Damanhuri, N. A. Othman, M. H. Othman, and W. F. A. W. Zaidi, “Potato leaf disease classification using image processing and artificial neural net-work,” in Proc. 2022 IEEE 12th Int. Conf. Control Syst., Comput. Eng. (ICCSCE), 2022, pp. 107–112, doi: 10.1109/ICCSCE54767.2022.9935654.
[19] A. Baranwal, M. Mishra, A. Goyal, and Yogesh, “Potato plant disease classification through deep learning,” in Proc. 2022 Int. Conf. Mach. Learn., Big Data, Cloud Parallel Comput. (COM-IT-CON), 2022, pp. 673–681, doi: 10.1109/COM-IT-CON54601.2022.9850701.
[20] K. El Moutaouakil, B. Jabir, and N. Falih, “A convolutional neural networks-based ap-proach for potato disease classification,” in Business Intelligence (CBI 2022) (Lecture Notes in Business Information Processing). Cham, Switzerland: Springer, 2022, pp. 29–40, doi: 10.1007/978-3-031-06458-6_2.
[21] E. Liljeroth, L. Stridh, E. Edin, and R. S. Sajeevan, “Early blight in potato: Decade-long field studies show high variation in yield losses,” Potato Res., vol. 68, no. 4, pp. 4909–4930, 2025, doi: 10.1007/s11540-025-09940-3.
[22] M. Dutta, D. Gupta, Y. Gulzar, M. S. Mir, C. W. Onn, and A. B. Soomro, “Leveraging Incep-tion V3 for precise early and late blight disease classification in potato crops,” Trait. Sig-nal, vol. 41, no. 2, pp. 705–715, 2024, doi: 10.18280/ts.410213.
[23] R. Asakaviciute, A. Zelya, T. Andriychuk, and A. Razukas, “Evaluation of potato varieties for yield, quality, and late blight resistance,” Life, vol. 15, no. 9, p. 1378, 2025, doi: 10.3390/life15091378.
[24] J. Rashid, I. Khan, G. Ali, S. H. Almotiri, M. A. AlGhamdi, and K. Masood, “Multi-level deep learning model for potato leaf disease recognition,” Electronics, vol. 10, no. 17, p. 2064, 2021, doi: 10.3390/electronics10172064.
[25] D. P. Hughes and M. Salathé, “An open access repository of images on plant health to ena-ble the development of mobile disease diagnostics,” arXiv preprint arXiv:1511.08060, 2015, doi: 10.48550/arXiv.1511.08060.
[26] V. W. de Vargas et al., “Imbalanced data preprocessing techniques for machine learning: A systematic mapping study,” Knowl. Inf. Syst., vol. 65, no. 1, pp. 31–57, 2023, doi: 10.1007/s10115-022-01772-8.
[27] A. Tawakuli, B. Havers, V. Gulisano, D. Kaiser, and T. Engel, “Survey: Time-series data pre-processing: A survey and an empirical analysis,” J. Eng. Res., vol. 13, no. 2, pp. 674–711, 2025, doi: 10.1016/j.jer.2024.02.018.
[28] K. M. Rao, S. Ghanta, and K. Supriya, “Data preprocessing techniques: Emergence and se-lection towards machine learning models – a practical review using HPA dataset,” Mul-timed. Tools Appl., vol. 82, no. 24, pp. 37177–37196, 2023, doi: 10.1007/s11042-023-15087-5.
[29] G. Verma, “Eye disease classification enhanced by EfficientNetB3: A deep learning ap-proach,” in Proc. 2024 Int. Conf. Artif. Intell. Emerg. Technol. (Global AI Summit), 2024, pp. 1016–1021, doi: 10.1109/GlobalAISummit62156.2024.10947971.
[30] M. T. Ramakrishna et al., “Leveraging EfficientNetB3 in a deep learning framework for high-accuracy MRI tumor classification,” Comput. Mater. Contin., vol. 81, no. 1, pp. 867–883, 2024, doi: 10.32604/cmc.2024.053563.
[31] A. Amato and V. Di Lecce, “Data preprocessing impact on machine learning algorithm per-formance,” Open Comput. Sci., vol. 13, no. 1, p. 20220278, 2023, doi: 10.1515/comp-2022-0278.
[32] B. Bala and S. Behal, “A brief survey of data preprocessing in machine learning and deep learning techniques,” in Proc. 2024 8th Int. Conf. I-SMAC (IoT Soc., Mobile, Anal. Cloud), 2024, pp. 1755–1762, doi: 10.1109/I-SMAC61858.2024.10714767.
[33] J. Hu, “High-resolution processing and feature extraction of logging curves based on wave-let transform and principal component analysis,” in Proc. 2025 IEEE 4th Int. Conf. Safe Prod. Informatiz. (IICSPI), 2025, pp. 389–397, doi: 10.1109/IICSPI66775.2025.11437785.
[34] C. E. Emere, C. O. Aigbavboa, W. D. Thwala, et al., “A principal component analysis of sus-tainable building construction features for project delivery in South Africa,” J. Eng. Des. Technol., vol. 23, no. 5, pp. 1463–1482, 2025, doi: 10.1108/JEDT-01-2024-0015.
[35] I. T. Jolliffe and J. Cadima, “Principal component analysis: A review and recent develop-ments,” Philos. Trans. R. Soc. A, vol. 374, no. 2065, p. 20150202, 2016, doi: 10.1098/rsta.2015.0202.
[36] S. Lloyd, “Least squares quantization in PCM,” IEEE Trans. Inf. Theory, vol. 28, no. 2, pp. 129–137, 1982, doi: 10.1109/TIT.1982.1056489.
[37] A. K. Jain, “Data clustering: 50 years beyond K-means,” Pattern Recognit. Lett., vol. 31, no. 8, pp. 651–666, 2010, doi: 10.1016/j.patrec.2009.09.011.
[38] O. Kisi, S. Heddam, K. S. Parmar, A. Petroselli, C. Külls, and M. Zounemat-Kermani, “Inte-gration of Gaussian process regression and K means clustering for enhanced short term rainfall runoff modeling,” Sci. Rep., vol. 15, no. 1, p. 7444, 2025, doi: 10.1038/s41598-025-91339-8.
[39] G. Verma, K. S. Kumar, N. Sharma, S. Kaparwan, and R. Madan, “Optimized EfficientNetB3 framework for high-precision apple leaf disease detection and classification,” in Proc. 2026 Contemp. Comput. Innov. Conf. (CCIC), 2026, pp. 1–6, doi: 10.1109/CCIC68129.2026.11485912.
[40] M. Velamakanni, S. S. Kanumalli, A. P. Gopi, V. L. Narayana, V. Sujatha, and A. Naresh, “Tomato leaf disease detection using EfficientNetB3,” in Proc. 2025 IEEE Madhya Pradesh Sect. Conf. (MPCON), 2025, pp. 765–771, doi: 10.1109/MPCON66082.2025.11256726.
[41] B. Bataineh et al., “Explainable hybrid deep learning with EfficientNetB3–DenseNet121 feature fusion for banana leaf spot disease classification,” Int. J. Adv. Signal Image Sci., vol. 12, no. 1, pp. 1120–1137, 2026, doi: 10.29284/yeveds27.
[42] E. S. Mughele, S. C. Chiemeke, and J. D. Ebube, “Tomato leaf disease detection using trans-fer learning with EfficientNetB3: A CNN-based approach,” Caliphate J. Sci. Technol., vol. 7, no. 3, pp. 454–459, 2025, doi: 10.4314/cajost.v7i3.18.
[43] H. Drucker, C. J. C. Burges, L. Kaufman, A. Smola, and V. Vapnik, “Support vector regres-sion machines,” in Advances in Neural Information Processing Systems 9. Cambridge, MA, USA: MIT Press, 1997, pp. 155–161.
[44] X. Yang, P. Meng, Z. Jiang, and L. Zhou, “Deep siamese residual support vector machine with applications to disease prediction,” Comput. Biol. Med., vol. 196, p. 110693, 2025, doi: 10.1016/j.compbiomed.2025.110693.
[45] J. Rybak, H. Battey, and W.-X. Zhou, “On inference for the support vector machine,” J. Mach. Learn. Res., vol. 26, no. 85, pp. 1–54, 2025. [Online]. Available: https://jmlr.org/papers/v26/23-1581.html
[46] W. Zheng, X. Yuan, X. Bao, and Y. Dong, “Adaptive support vector machine for time-variant failure probability function estimation,” Reliab. Eng. Syst. Saf., vol. 253, p. 110510, 2025, doi: 10.1016/j.ress.2024.110510.
[47] Z. Mohamadi et al., “The application of random forest-based models in prognostication of gastrointestinal tract malignancies: A systematic review,” Front. Artif. Intell., vol. 8, p. 1517670, 2025, doi: 10.3389/frai.2025.1517670.
[48] R. O. Ogundokun, P. A. Owolawi, and E. A. van Wyk, “A lightweight deep feature fusion and PCA-optimized random forest framework for multi-class potato leaf disease classification,” in Applied Informatics (Communications in Computer and Information Science). Cham, Switzerland: Springer, 2025, pp. 147–162, doi: 10.1007/978-3-032-07175-0_10.
[49] J. Korstanje, “The random forest,” in Advanced Forecasting with Python, 2nd ed. Berkeley, CA, USA: Apress, 2025, pp. 191–204, doi: 10.1007/979-8-8688-2028-1_15.
[50] B. Mallala, A. I. U. Ahmed, S. V. Pamidi, M. O. Faruque, and R. Reddy M, “Forecasting global sustainable energy from renewable sources using random forest algorithm,” Results Eng., vol. 25, p. 103789, 2025, doi: 10.1016/j.rineng.2024.103789.
[51] B. Tamiru, T. Soromessa, B. Warkineh, and G. Legese, “Modelling selected soil chemical properties using TerrSet and random forest: A case of Hangadi watershed, Oromia, Ethio-pia,” Geol. Ecol. Landscapes, vol. 9, no. 4, pp. 1177–1193, 2025, doi: 10.1080/24749508.2024.2429842.
[52] Y. Xia, W. Lu, Z. Peng, J. Lou, J. Huang, and J. Niu, “Understanding the impacts of space design on local outdoor thermal comfort: An approach combining DepthmapX and XGBoost,” Energy Build., vol. 332, p. 115451, 2025, doi: 10.1016/j.enbuild.2025.115451.
[53] Y. Liu et al., “Urban flood depth prediction and visualization based on the XGBoost-SHAP model,” Water Resour. Manag., vol. 39, no. 3, pp. 1353–1375, 2025, doi: 10.1007/s11269-024-04020-6.
[54] M. Imani, A. Beikmohammadi, and H. R. Arabnia, “Comprehensive analysis of random for-est and XGBoost performance with SMOTE, ADASYN, and GNUS under varying imbalance levels,” Technologies, vol. 13, no. 3, p. 88, 2025, doi: 10.3390/technologies13030088.
[55] H. Zhuo, T. Li, W. Lu, Q. Zhang, L. Ji, and J. Li, “Prediction model for spontaneous combus-tion temperature of coal based on PSO-XGBoost algorithm,” Sci. Rep., vol. 15, no. 1, p. 2752, 2025, doi: 10.1038/s41598-025-87035-2.
Downloads
Published
License
Copyright (c) 2026 Bimantyoso Hamdikatama, Kusrini, Arif Setyanto, Julius Jeremiah Allieu
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
