Related Experiment Video
Updated: Jun 17, 2025
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
380
AISOA-SSformer: An Effective Image Segmentation Method for Rice Leaf Disease Based on the Transformer Architecture
Weisi Dai1, Wenke Zhu2, Guoxiong Zhou1
1Faculty of Electronic Information and Physics, Central South University of Forestry and Technology, Changsha, 410004 Hunan, China.
Plant Phenomics (Washington, D.C.)
|August 6, 2024
Summary
This study introduces AISOA-SSformer, a Transformer-based algorithm for accurate rice leaf disease segmentation. It enhances feature extraction and model robustness, improving disease identification for modern farming.
Area of Science:
- Agricultural Science
- Computer Vision
- Machine Learning
Background:
- Rice leaf diseases significantly impact crop yield and food security.
- Accurate disease identification is vital for effective crop management.
- Existing segmentation methods face challenges due to environmental diversity and disease complexity.
Purpose of the Study:
- To develop an innovative semantic segmentation algorithm for rice leaf pests and diseases.
- To enhance the accuracy and robustness of disease identification in rice cultivation.
- To provide farmers with advanced tools for modern plantation management.
Main Methods:
- Introduced AISOA-SSformer, a Transformer-based semantic segmentation algorithm.
- Implemented a sparse global-update perceptron for real-time parameter updating.
- Utilized a salient feature attention mechanism with spatial (SRM) and channel (CRM) reconstruction modules.
- Employed an annealing-integrated sparrow optimization algorithm for fine-tuning.
Main Results:
- AISOA-SSformer achieved 83.1% MIoU, 80.3% Dice coefficient, and 76.5% recall on a custom dataset.
- The model boasts a compact size of 14.71 million parameters.
- Demonstrated superior accuracy in rice leaf disease segmentation compared to existing algorithms.
Conclusions:
- AISOA-SSformer effectively improves rice leaf disease segmentation accuracy and robustness.
- The developed method offers valuable insights for precision agriculture and disease management.
- Open-sourced dataset and code facilitate further research and application.
