Reproduce “Optimized Custom CNN for Real-Time Tomato Leaf Disease Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Mangsura Kabir Oni, Tabia Tanzin Prama
An optimized custom Convolutional Neural Network (CNN) achieved 95.2% accuracy in real-time binary detection of healthy versus diseased tomato leaves using a localized dataset from Bangladesh. This custom model surpassed the performance of established architectures like MobileNetV2, YOLOv5, and ResNet18 for this specific task.
Key insights
- Tailoring CNN architectures to specific datasets and problem statements can yield superior performance compared to generic pre-trained models.
- Utilizing localized datasets is crucial for developing robust and practically applicable disease detection systems that reflect regional environmental conditions and disease profiles.
- A carefully optimized custom model can achieve high accuracy in a critical agricultural task like disease detection, even outperforming powerful off-the-shelf models.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.