KNUST_Eccentric Griffins

An android AI application that detects tomato leaf disease and provides suggestion to minimize the effect of the detected cause. The application has been translated in Twi, a native Ghanaian language

Comment

Inspiration

Bringing an AI solution to all the stakeholders regardless of their social and economic status.

What it does

How we built it

About the project:

Our solution software detects plant diseases from just a click! Our Application contributes to mitigating a major factor that can cause economy failure of states across the globe. This is an aid to farmers so crop diseases can be detected early and proper steps can be taken to maximize yield. We also provide audio support for farmers with weak vision.

Built with: Python, PyTorch, Google Colab, Kaggle, Roboflow, Android Studio, Google Translate API, Canva Prototype Try it out links: https://github.com/QasimKhan5x/tumhackathon

Challenges we ran into

  • Issues with model inference in pytorch vs android
  • Building the frontend because we don't have requisite experience
  • Unable to deploy application and integrate detection and classification

Accomplishments that we're proud of

  • 95% F1 Score with lightweight model
  • Created largest tomato leaf dataset as of 2022

- Being able to deploy offline end-to-end application without prior experience

What we learned

  • Android development
  • Time management
  • Teamwork
  • Extra skills for learning

What's next for KNUST_Eccentric Griffins

  • Add support for cocoa beans detection
Share this project:

Updates