Related Experiment Videos
ZDAM: a new deep learning model for bean leaf disease diagnosis
Jia Liu1,2, Kaidi Yu1, Hongyun Song1
1School of Computer and Artificial Intelligence, Henan University of Urban Construction, Pingdingshan, China.
Frontiers in Plant Science
|June 29, 2026
Summary
A new deep learning model, ZDAM, accurately identifies crop diseases from leaf images, achieving 99.02% accuracy. This automated disease monitoring supports sustainable agriculture and reduces postharvest losses.
Area of Science:
- Agricultural Science
- Computer Science
- Plant Pathology
Background:
- Accurate crop disease diagnosis is vital for agricultural productivity and food safety.
- Traditional methods struggle with complex field conditions and extensive feature modeling.
Purpose of the Study:
- To develop an advanced deep learning model for precise crop disease identification.
- To improve automated disease monitoring systems for enhanced agricultural management.
Main Methods:
- Proposed a deep learning model, ZDAM, utilizing an optimized ZFNet with a dual attention mechanism.
- Integrated channel and spatial attention, along with a residual module, to refine feature extraction and boost accuracy.
Main Results:
- The ZDAM model achieved an average recognition accuracy of 99.02% on a dataset of 11,903 bean leaf images.
- Outperformed existing models like MobileMamba, Vision Transformer, and Chest-OMD in disease identification.
Conclusions:
- The ZDAM model provides a scalable and accurate solution for automated crop disease monitoring.
- This technology aids in preserving postharvest quality and promoting sustainable crop production.
