A Convolutional Neural Network (CNN) project was completed as part of my machine learning course in Fall 2022. This project aimed to prove the concept that applying CNN to plant leaves can accurately identify disease. The project focused on tomato leaves and now needs to be applied to hop leaves. The file named ‘Hops_CNN’ contains the model and a tutorial on importing, cleaning, and formatting the data as well as applying the model to the data. The tutorial includes links to external resources for further study and explanation, and the project draws from several sources of methodology. The model achieved over 90% accuracy in both training and testing, with a decreasing loss over epochs.
- Hops_CNN.ipynb - Python notebook originally written in Google Colab containing the code and tutorial for this project
- Tomato_images.zip - compressed folder containing three folders of tomato image data
- Poster_Hops_CNN.pdf - poster created for the project overview, support, and findings presented during the class
- Abstract.md - abstract on overall project, aimed at applying model to images of hop plants
- CNN_References.bib - contains references for articles and sources used in the Hops_CNN.ipynb notebook
This project was developed, and intended to run, using Google Colab. Below are the directions for downloading the notebook and data set to run this code.
- Download the Hops_CNN.ipynb and Tomato_Images.zip files from this repository.
- Create a new folder on your local machine and move the Hops_CNN and Tomato_Images files into that folder. Unzip the Tomato_Images folder and delete the zipped copy.
- Upload the folder containing the Hops_CNN and Tomato_Images files to your Google Drive.
- Once uploaded, navigate to the folder now in your Google Drive and open the Hops_CNN.ipynb using Google Colab.
- Follow the tutorial in the notebook for mounting the data to your drive and executing the code.