Earticle

초록

영어

This study presents a lesion-guided four-channel ConvNeXt model for tomato plant disease recognition. By segmenting lesion areas in the HSV color space, grayscale masks are generated and combined with RGB images to form a four-channel input. The proposed ConvNeXt4Channel network, optimized for this input, enhances spatial feature extraction. Experiments on the PlantVillage tomato dataset (train-test split: 8:2) show that the model, trained with cross-entropy loss and Adam optimizer (learning rate = 1e-4), achieves 96.81% accuracy—surpassing conventional models by approximately 2.5%. Grad-CAM visualizations indicate improved lesion localization, confirming the effectiveness of lesion-guided enhancement. This method provides a robust and interpretable solution for automated crop disease diagnosis.

목차

Abstract
1. Introduction
2. Related Work
3. Methodologies
3.1 Disease Spot Region Extraction
3.2 Construction of Four-Channel Input
3.3 Training and Optimization Strategy
4. Experiments
4.1 Experimental Settings
4.2 Overall Performance
4.3 Comparison with Other Models
4.4 Grad-CAM Visualization Analysis
5. Conclusion
Acknowledgement
References

저자

• Yue Zhang [ Department of Computer Science and Engineering, Sejong University ]
• Ji-Won Kim [ Department of Computer Science and Engineering, Sejong University ]
• L. Minh Dang [ Department of Computer Science and Engineering, Sejong University ]
• Wenqi Zhang [ Department of Computer Science and Engineering, Sejong University ]
• Yanan Wang [ Department of Computer Science and Engineering, Sejong University ]
• Xingshi Gan [ Department of Computer Science and Engineering, Sejong University ]
• Hyeonjoon Moon [ Department of Computer Science and Engineering, Sejong University ] Corresponding Author

참고문헌