/Finetuning-with-Pre_Trained-Mobilenet_v2-for-Crop-Disease-Classification---Agriculture-Domain

Utilize mobilenet_v2 and Finetune it with potato disease image dataset (3 classes). App will allow farmers to snap a picture of a plant and determine whether the plant has a disease or not.

Primary LanguageJupyter NotebookGNU General Public License v3.0GPL-3.0

Project Info

This is a deep learning project in the field of agriculture.

Each year, farmers suffer economic losses and crop loss as a result of different illnesses. We will utilize mobilenet_v2 Convolutional Neural Network which is pre-trained on the ImageNet dataset and Finetune it with plant disease image dataset. The goal is to classify images and develop an app/mobile app that will allow farmers to snap a picture of a plant and the app will determine whether the plant has a disease or not.

Here we'll be doing potato disease classification.

  • Dataset consist of leaf images from healthy potato plants and diseased potato plants (Early blight and Late blight)
  • 3 classes
    • Early blight, Late blight, Healthy

This project's technology stack will include the following:

  • Tensorflow with Keras API
  • TF dataset
  • Data Augmentation
  • mobilenet_v2 model pre-trained on the ImageNet dataset.
  • Tensorflow Lite (Useful for mobile dev)