The AI Pipeline
How PlantCare AI Works
Our AI system follows a deep learning pipeline to detect plant diseases accurately.
1
Image Upload
You upload a clear image of the affected plant leaf.
2
Preprocessing
The image is resized, normalized, and prepared using the same pipeline used during training.
3
EfficientNetB0
Our model, pretrained on ImageNet and fine-tuned on plant-disease imagery, extracts important features from the image.
4
AI Prediction
The model predicts the most likely disease class with a confidence score.
5
Grad-CAM
Grad-CAM highlights the image regions that most influenced the AI's prediction.
Why EfficientNetB0?
EfficientNetB0 balances accuracy and computational cost well, making it a practical choice for CPU-oriented deployment.
High Accuracy
Lightweight
Fast Inference
Transfer Learning