Inspiration
The inspiration for this project came from a high school class where we learned about a devastating crop disease in China that wiped out thousands of plants. The farmers had never encountered this disease before, making it difficult to identify and manage in time. This sparked my desire to develop a solution that could help farmers quickly and accurately identify plant diseases.
What it does
My app, built in Xcode, uses a ResNet50 computer vision model trained through transfer learning. Users can upload a photo of a plant from their camera roll, and the model identifies the plant and detects any diseases it might have. This could be invaluable for early disease detection and crop protection.
How we built it
To build the app, I utilized a publicly available dataset from Kaggle and applied transfer learning on the ResNet50 model. I then converted the model to a format compatible with Xcode. The app integrates with the camera functionality on iOS devices, enabling users to take or upload photos directly for analysis.
Challenges we ran into
One major challenge was converting the model into Apple’s specific machine learning format, which required careful handling. Additionally, integrating the model with the camera function in Xcode was complex due to the permissions required for camera access.
Accomplishments that we're proud of
I'm proud that it works and that I created something that can help people.
What we learned
Through this project, I learned how to make machine learning models lightweight and portable enough to be used on mobile devices, making advanced technology accessible to anyone, anywhere.
What's next for PlantDiseaseClassifier
Moving forward, I plan to further refine the model’s accuracy and consider publishing the app to make it widely available.
Log in or sign up for Devpost to join the conversation.