An Empirical Study of Hyperparameter Impact on Deep Learning Models for Cardamom Leaf Disease Classification

Year : 2025 | Volume : 15 | Issue : 03 | Page : 48 60
    By

    Anjana Jayan,

  • Benson C.C.,

  1. Research Scholar, Department of Computer Science, St. Thomas College (Autonomous), Thrissur, University of Calicut, Kerala, India
  2. Assistant Professor, Department of Computer Application, Mar Dionysius College, Pazhanji, Thrissur, Kerala, India

Abstract

Recent advancements in deep learning models like convolutional neural networks and self- attention mechanisms have achieved great success in the field of plant disease classification. This study investigates the efficacy of two pre-trained models, ConvNeXT-Tiny and Swin Transformer-Tiny, for leaf disease classification in cardamom using a publicly available dataset constituting three categories of leaves, namely Healthy, Colletotrichum Blight and Phyllosticta Leaf Spot. The effectiveness of the models highly depends on appropriate selection of hyperparameters. The paper presents a comprehensive evaluation of the performance of models in terms of different combinations of learning rate, batch size, optimizers and number of epochs. The experimental results demonstrated that the ConvNeXT-Tiny model performed better than the Swin Transformer-Tiny, achieving a higher accuracy of 99.23%, for an optimal configuration of hyperparameters. This research emphasizes that the right choice of hyperparameters can make a significant difference to the effectiveness of systems designed to classify leaf diseases.

Keywords: Deep Learning, ConvNeXT, Swin Transformer, Hyperparameters, Cardamom Leaf Disease Classification

[This article belongs to Current Trends in Information Technology ]

How to cite this article:
Anjana Jayan, Benson C.C.. An Empirical Study of Hyperparameter Impact on Deep Learning Models for Cardamom Leaf Disease Classification. Current Trends in Information Technology. 2025; 15(03):48-60.
How to cite this URL:
Anjana Jayan, Benson C.C.. An Empirical Study of Hyperparameter Impact on Deep Learning Models for Cardamom Leaf Disease Classification. Current Trends in Information Technology. 2025; 15(03):48-60. Available from: https://journals.stmjournals.com/ctit/article=2025/view=232598


References

  1. Vijayan AK. Small cardamom production technology and future prospects. International Journal of Agriculture Sciences, ISSN. 2018:0975-3710.
  2. Sunil CK, Jaidhar CD. Cardamom plant disease detection approach using EfficientNetV2. Ieee Access. 2021 Dec 27;10:789-804.
  3. Shewale MV, Daruwala R. Impact of hyperparameter tuning for identification and classification of plant leaf diseases: a deep learning approach. In2022 IEEE Conference on Interdisciplinary Approaches in Technology and Management for Social Innovation (IATMSI) 2022 Dec 21 (pp. 1- 5). IEEE.
  4. Liu Z, Mao H, Wu CY, Feichtenhofer C, Darrell T, Xie S. A convnet for the 2020s. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition 2022 (pp. 11976-11986).
  5. Liu Z, Lin Y, Cao Y, Hu H, Wei Y, Zhang Z, Lin S, Guo B. Swin transformer: Hierarchical vision transformer using shifted windows. InProceedings of the IEEE/CVF international conference on computer vision 2021 (pp. 10012-10022).
  6. Alam TS, Jowthi CB, Pathak A. Comparing pre-trained models for efficient leaf disease detection: a study on custom CNN. Journal of Electrical Systems and Information Technology. 2024 Feb 23;11(1):12.
  7. Vats S, Singh AN, Kukreja V, Sharma R. Leveraging pre-trained deep learning models for orange leaf disease classification. In2024 IEEE 9th International Conference for Convergence in Technology (I2CT) 2024 Apr 5 (pp. 1-4). IEEE.
  8. Hukkeri GS, Soundarya BC, Gururaj HL, Ravi V. Classification of various plant leaf disease using pretrained convolutional neural network on imagenet. The Open Agriculture Journal. 2024 May 20;18(1).
  9. Chellapandi B, Vijayalakshmi M, Chopra S. Comparison of pre-trained models using transfer learning for detecting plant disease. In2021 International Conference on Computing, Communication, and Intelligent Systems (ICCCIS) 2021 Feb 19 (pp. 383-387). IEEE.
  10. Rastogi P, Dua S, Dagar V. Early Disease Detection in Plants using CNN. Procedia Computer Science. 2024 Jan 1;235:3468-78.
  11. Bezabh YA, Salau AO, Abuhayi BM, Mussa AA, Ayalew AM. CPD-CCNN: Classification of pepper disease using a concatenation of convolutional neural network models. Scientific Reports. 2023 Sep 20;13(1):15581..
  12. Yan J, Mo Y, Yu Y, Dou S, Yang R. Citrus disease classification model based on improved ConvNeXt. IEEE Access. 2024 Jun 20;12:152498-510..
  13. KP AR, Gowrishankar S. Convnext-based mango leaf disease detection: Differentiating pathogens and pests for improved accuracy. International Journal of Advanced Computer Science and Applications. 2023;14(6)..
  14. Wang X, Wang Y, Zhao J, Niu J. Eca-convnext: A rice leaf disease identification model based on convnext. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition 2023 (pp. 6235-6243)..
  15. Li H, Qi M, Du B, Li Q, Gao H, Yu J, Bi C, Yu H, Liang M, Ye G, Tang Y. Maize disease classification system design based on improved ConvNeXt. Sustainability. 2023 Oct 13;15(20):14858..
  16. Bajpai A, Tiwari N, Rajput P, Sahu S, Singh D. Enhanced potato leaf disease detection via modified swin transformer architecture. In2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT) 2024 Jun 24 (pp. 1-7). IEEE..
  17. Sarkar S, Kumar G. Identification of sugarcane leaf diseases and deficiency disorders using transformers. In2024 First International Conference on Pioneering Developments in Computer Science & Digital Technologies (IC2SDT) 2024 Aug 2 (pp. 493-498). IEEE..
  18. Li LH, Tanone R. Disease identification in potato leaves using swin transformer. In2023 17th International Conference on Ubiquitous Information Management and Communication (IMCOM) 2023 Jan 3 (pp. 1-5). IEEE..
  19. Chavarro AF, Renza D, Ballesteros DM. Influence of hyperparameters in deep learning models for coffee rust detection. Applied Sciences. 2023 Apr 4;13(7):4565..
  20. Gunarathna MM, Rathnayaka RM. Experimental determination of CNN hyper-parameters for tomato disease detection using leaf images. In2020 2nd international conference on advancements in computing (ICAC) 2020 Dec 10 (Vol. 1, pp. 464-469). IEEE..
Regular Issue Subscription Review Article
Volume 15
Issue 03
Received 08/08/2025
Accepted 10/09/2025
Published 17/09/2025
Publication Time 40 Days
148

Login


My IP

PlumX Metrics