Research on the intelligent detection model of plant diseases based on MamSwinNet
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.A new MamSwinNet model efficiently detects plant diseases using refined tokens and global perception modules. This advanced AI solution improves accuracy and significantly reduces computational costs for better crop security.
Area Of Science
- Agricultural Science
- Computer Vision
- Artificial Intelligence
Background
- Plant diseases significantly threaten global crop yield, quality, and food security.
- Existing Convolutional Neural Network (CNN) and Transformer-based methods have limitations in modeling dependencies, generalization, and computational efficiency.
- Accurate and efficient plant disease detection is crucial for sustainable agriculture.
Purpose Of The Study
- To propose the MamSwinNet model, an innovative deep learning architecture for efficient and accurate plant disease detection.
- To address the limitations of existing CNN and Transformer models in handling long-range dependencies and computational complexity.
- To provide a reliable intelligent solution for large-scale crop disease detection.
Main Methods
- Developed MamSwinNet incorporating an Efficient Token Refinement module with overlapping space reduction using depthwise separable convolutions.
- Integrated Spatial Global Selective Perception (SGSP) and Channel Coordinate Global Optimal Scanning (CCGOS) modules for enhanced feature modeling.
- Utilized 2D-SSM for long-range dependency capture and Mamba blocks for channel-selective attention.
Main Results
- MamSwinNet achieved high F1 scores: 79.47% (PlantDoc), 99.52% (PlantVillage), and 99.38% (Cotton).
- The model boasts 12.97M parameters (52.9% less than Swin-T) and a computational cost of 2.71GMac.
- Demonstrated significant improvements in computational efficiency and feature retention compared to existing models.
Conclusions
- MamSwinNet offers an efficient and reliable intelligent solution for plant disease detection.
- The proposed architecture effectively balances accuracy and computational efficiency.
- This work contributes to advancing AI applications in agriculture for enhanced food security.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
