Related Experiment Video
Updated: May 16, 2025
08:47
Author Spotlight: UAV Remote Sensing for Efficient Invasive Plant Biomass Estimation
Published on: February 9, 2024
1.3K
A compact deep learning approach integrating depthwise convolutions and spatial attention for plant disease
Amreen Batool1, Jisoo Kim2, Yung-Cheol Byun3
1Department of Electronic Engineering Institute of Information Science & Technology, Jeju National University, Jeju, 63243, Korea.
Plant Methods
|April 2, 2025
Summary
A new lightweight deep learning model (LWDSC-SA) accurately detects plant leaf diseases. This efficient model improves crop health management in resource-limited agricultural settings.
Area of Science:
- Agricultural Science
- Computer Science
- Machine Learning
Background:
- Plant diseases pose significant threats to agricultural productivity and global food security.
- Existing deep learning models struggle with intricate leaf features and are computationally expensive for farm deployment.
Purpose of the Study:
- To develop a lightweight deep learning model for efficient and accurate plant disease detection.
- To enhance feature extraction capabilities for improved classification of plant leaf diseases.
Main Methods:
- Proposed a novel Lightweight Deep Learning model with Depthwise Separable Convolution and Spatial Attention (LWDSC-SA).
- Evaluated the model on the PlantVillage dataset (38 classes, 55,000 images).
- Utilized K-fold cross-validation (K=5) to assess robustness and generalizability.
Main Results:
- Achieved 98.7% accuracy on the PlantVillage dataset, outperforming MobileNet, MobileNetV2, AlexNet, and VGGNet16.
- Demonstrated high average performance with K-fold cross-validation: 98.58% accuracy, 98.30% precision, 98.90% recall, and 98.58% F1 score.
- The LWDSC-SA model proved computationally efficient and effective in feature extraction.
Conclusions:
- The LWDSC-SA model offers a superior, lightweight, and efficient solution for plant disease detection.
- Enables effective crop health management in resource-constrained agricultural environments.
- Contributes to sustainable agricultural practices through accurate and accessible disease identification.
