Plant disease recognition using residual convolutional enlightened Swin transformer networks.

Area of Science:

• Agricultural Science
• Computer Science
• Machine Learning

Background:

• Plant diseases significantly impact agricultural economies and global food security.
• Early detection of plant diseases is crucial for effective management and yield preservation.
• Traditional methods struggle with the complexity and subtle variations in plant disease symptoms.

Purpose of the Study:

• To develop a highly accurate deep learning model for early-stage plant disease prediction.
• To address challenges in precise disease identification, including noise, intensity variations, and visual similarities.
• To improve upon existing hybrid learning models for plant disease detection.

Main Methods:

• Proposed a novel ensemble of Swin transformers (ST) and residual convolutional networks.
• Utilized Swin transformers for hierarchical feature extraction and residual networks for deep key-point feature extraction.
• Conducted extensive experiments on the Plant Village Kaggle dataset, evaluating accuracy, precision, recall, specificity, and F1-rating.

Main Results:

• The proposed Swin transformer and residual network ensemble demonstrated superior performance compared to state-of-the-art models.
• Achieved higher accuracy, precision, recall, and F1-rating in plant disease identification.
• The model effectively handled complex variations and noise in leaf imagery.

Conclusions:

• The novel ensemble deep learning model offers a significant advancement in automated plant disease detection.
• This approach provides a robust solution for early and precise identification of plant diseases.
• The findings support the integration of advanced AI techniques in precision agriculture for enhanced crop management.