One of the most overlooked jobs in our modern world is our farmers, who, through sweat and hard work, are the foundation of our century's exponential intellectual and technological growth. We wanted, through our project, to give back to those who have served us so much, and who filled our plates with quality food. With the increasing prices and the shortcoming of production due to the lack of labor in recent years, we wished to look for a solution to help farmers deal with their crops. We present Lifting.
What it does
Through an image recognition program founded upon a convolutional neural network, Leafting can detect the presence of diseases in crops, while developing solutions to help farmers act before it gets too late.
How we built it
Flask, Python, HTML5, CSS for the front-end Numpy, Pandas, OpenCV, Keras, Scikit-learn for backend
Challenges we ran into
The many classes present in our model led to a high run-time. We had to use optimization techniques for our model. We encountered difficulties exporting the model to a single function and relaying the input image from the website to the model.
Accomplishments that we're proud of
We are proud of the efficiency and accuracy of the model. The website is simple and easy to use.
What we learned
We learned Flask, TensorFlow and CNN, and exporting models.
What's next for Leafting
Democratize our product and expand classes and datasets