TCSRNet: A Lightweight Model for Recognizing Tobacco Leaf Curing Stages
Author Information
Author(s): Zhao Panzhen, Wang Songfeng, Duan Shijiang, Wang Aihua, Meng Lingfeng, Hu Yichong
Primary Institution: Tobacco Research Institute of Chinese Academy of Agricultural Sciences
Hypothesis
Can a lightweight classification network model effectively recognize tobacco leaf curing stages while balancing accuracy and computational efficiency?
Conclusion
The TCSRNet model achieved a classification accuracy of 90.3% with significantly reduced computational complexity, making it suitable for real-time applications in resource-constrained environments.
Supporting Evidence
- TCSRNet achieved a classification accuracy of 90.35% on the tobacco leaf curing stage dataset.
- The model has a computational complexity of only 158.136 MFLOPs and a parameter count of 1.749M.
- TCSRNet outperformed other models like ResNet34 and MobileNetV3 in terms of accuracy and efficiency.
- When tested on the V2 Plant Seedlings dataset, TCSRNet maintained an accuracy of 97.15%.
Takeaway
This study created a smart model that helps farmers know when their tobacco leaves are ready by looking at pictures, making it easier to grow good tobacco.
Methodology
The study developed TCSRNet using an Inception structure, Ghost modules, and a Multi-scale Adaptive Attention Module to enhance feature extraction and reduce computational complexity.
Limitations
The model's performance may vary across different datasets and regions, and real-world validation in embedded systems is still needed.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website