Plant-Disease-Prediction-CNN

A CNN model to classify different plant diseases. Vgg16 net is fine tuned to the crowdAI dataset. The dataset can be found at https://www.crowdai.org/challenges/1

Requirements

• python 2.7
• numpy
• tensorflow
• keras
• OpenCV (can use pillow Library Too)

Methodology

The pre-trained VGG16 model trained on imagenet dataset was used and the last 4 convolutional layers were trained on our dataset. This process is called fine tuning. A point to note is that only 19 classes were used from the dataset since we were focussing on a particular region where only these crops are grown.