The use of AI ML to predict plant disease through image recognition and classification of plant leaves was the second stage in the AIML journey. The project was environment related, so it was personal to me, given to the current severe imbalance in our bio-spheres.
Instead of enlisting each and every step of the project such as -
data analysis, loading data into input layer, Processing in the middle layer such as flattening image into numpy binary array, loading array to prediction model and saving the prediction train as a serialized object to be used for display as in above image, while using Flask and wsgi web server,
I am herewith giving screenshots of my code.
Following code shows the libraries imported and the verification of dataset sanity.
2. Following code shows the creation of the CNN model and various hidden layers in the model.
3. Following code-snippet is self-explanatory on how it does iteration to re-enforce the training to CNN and improve prediction accuracy.
4. Plotting model accuracy. Please Seaborne's sns.distplot is deprecated.
5. Testing with Minimal Test data.
