Optimized Vision Transformers for Superior Plant Disease Detection
IEEE Access · 2025 · 13 · 48552-48570
DOI: 10.1109/access.2025.3547416 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2025. Verfasst von Abdelmalik Ouamane und 6 weiteren Autor:innen. Die Arbeit wurde bisher 37-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Detecting plant diseases is vital for maintaining agricultural productivity and ensuring food security. Advances in computer vision, particularly with Vision Transformers (ViTs), have shown significant potential in improving the accuracy and efficiency of plant disease identification. This study provides a comprehensive evaluation of various ViT parameters to determine the most effective configuration for this purpose. Using the extensive PlantVillage dataset, we systematically analyzed the effects of patch sizes, image resolutions, embedding dimensions, the number of transformer blocks (depth), the number of heads in the multi-head attention layer, and the dimension of the MLP (FeedForward) layer on model performance. We introduced saliency map visualizations to enhance interpretability and evaluate the critical regions contributing to classification decisions, ensuring the approach’s transparency and robustness. Our experiments identified the optimal ViT configuration as follows: image size$ = 224 \times 224$, patch size = 16, embedding dimension = 512, depth = 6, number of heads = 8, and MLP dimension = 1024. This configuration achieved an impressive accuracy of 99.77% on the PlantVillage dataset. In addition, we incorporated a novel cross-dataset transferability evaluation to validate the generalizability of the proposed model. Comparative analysis with traditional convolutional neural network architectures, such as VGG19 and AlexNet, revealed that our optimized ViT model not only surpasses these models in accuracy but also requires significantly fewer trainable parameters and storage space. The incorporation of a lightweight, domain-specific fine-tuning process ensures the model’s adaptability to new datasets with minimal computational overhead. Our findings highlight the scalability and adaptability of ViTs, emphasizing their ability to effectively handle varying image sizes and resolutions. Moreover, our approach outperforms recent state-of-the-art methods across multiple databases, underscoring the efficacy of the chosen ViT parameters.
Zitieren
APA
Ouamane, A., Chouchane, A., Himeur, Y., Miniaoui, S., Atalla, S., Mansoor, W., & Al-Ahmad, H. (2025). Optimized Vision Transformers for Superior Plant Disease Detection. IEEE Access, 13, 48552-48570. https://doi.org/10.1109/access.2025.3547416
Vancouver
Ouamane A, Chouchane A, Himeur Y, Miniaoui S, Atalla S, Mansoor W, et al. Optimized Vision Transformers for Superior Plant Disease Detection. IEEE Access. 2025;13:48552-48570. doi:10.1109/access.2025.3547416
BibTeX
@article{Ouamane2025,
author = {Ouamane, Abdelmalik and Chouchane, Ammar and Himeur, Yassine and Miniaoui, Sami and Atalla, Shadi and Mansoor, Wathiq and Al-Ahmad, Hussain},
title = {Optimized Vision Transformers for Superior Plant Disease Detection},
journal = {IEEE Access},
year = {2025},
volume = {13},
pages = {48552--48570},
doi = {10.1109/access.2025.3547416}
}
EUVIMED