Improved AlexNet with Inception-V4 for Plant Disease Diagnosis.

Area of Science:

• Agricultural Science
• Computer Science
• Machine Learning

Background:

• Manual plant disease identification is time-consuming and labor-intensive in large-scale agriculture.
• Accurate and timely disease detection is crucial for effective pest management and crop yield.
• Existing methods may lack the efficiency and accuracy required for modern agricultural challenges.

Purpose of the Study:

• To develop an efficient deep learning model for automated agricultural crop disease identification.
• To improve the accuracy and speed of disease detection compared to manual methods and existing models.
• To evaluate the model's performance on a diverse dataset of common crop diseases.

Main Methods:

• A novel deep learning architecture was proposed, combining and modifying AlexNet and Inception-V4.
• The model was trained and evaluated on an expanded PlantVillage dataset.
• Performance was benchmarked against established models including AlexNet, VGG11, Zenit, and VGG16.

Main Results:

• The proposed model achieved superior accuracy and F1 scores compared to all benchmarked methods.
• Highest accuracies were recorded for corn (94.5%), tomato (94.8%), grape (92.3%), and apple (96.5%).
• Correspondingly high F1 scores were obtained: corn (0.938), tomato (0.910), grape (0.945), and apple (0.924).

Conclusions:

• The developed deep learning model demonstrates significant potential for accurate and efficient crop disease identification.
• The model exhibits promising generalization capabilities across different crops and disease types.
• This approach offers a scalable solution to enhance disease management in modern agriculture.