AI
Plant Disease Detection
Computer-vision leaf-disease classifier via transfer learning, served in real time.
How it works
01 Leaf Image
Image inputA photo of a plant leaf is captured in the field for diagnosis.
96%
Accuracy
Real-time
Serving
Stack
TensorFlowTransfer LearningCNNTF ServingOpenCV
The Problem
Early plant-disease diagnosis in the field needs accuracy without huge labelled datasets.
Objective
Classify leaf diseases accurately and serve predictions for real-time field use.
Approach
Built a CNN with transfer learning and data augmentation to handle limited labelled data, then deployed on TensorFlow Serving for real-time diagnosis.
Challenges
Class imbalance across disease types; addressed with augmentation and transfer learning.
Results
96% detection accuracy on leaf-disease classification.
What's Next
On-device inference for offline field diagnosis.