Skip to content

Latest commit

 

History

History
24 lines (15 loc) · 1004 Bytes

File metadata and controls

24 lines (15 loc) · 1004 Bytes

Potato Disease Classification

This is a potato disease classification web app made using TensorFlow and Keras' Sequential API. It needs jpg image of a potato leaf and using Convolutional Neural Networks it predicts if the plant has:

  • Early Blight
  • Late Blight
  • No Disease (Healthy)

🔨 Model Architecture

The model is as shown below:

ModelArchitecture

⏳ Training

The model is trained on the dataset taken from Kaggle. In addition to the following dataset, data-augmentation was also performed to significantly increase the diversity of data available.

🚀 Deployment

The web app is deployed using StreamLit and Heroku. Here is the link to the app: Potato Blight Detector

Note that currently only jpg image is supported to give accurate predictions.