AgriFarm is an Android app for farmers to help diagnose crop diseases using online & on device offline machine learning
Our project helps farmers detect diseases of their crops offline because farmers tend to live in rural areas where internet doesn’t provide promising signals and farmers face hard time diagnosing their crops from apps that relies on internet connection. Our app also provides features that are dependent on internet connection such as weather, video recommendation based on farmer’s daily interactivity, prices prediction for the vegetables and fruits that are of highly important to farmers. Once we complete the disease detection module. We are also going to work on smart usage of limited internet connection (slow internet) to cache the important prices, news without farmers realizing glitch in performance. This is going to be a tough part of our project but this is where we need help of Google.
Currently, we have a customized trained model on Kaggle which is limited to few vegetables. We need Google’s help in order to expand the dataset to cover more fruits and vegetables beyond Rice, Onion, Bell Pepper, Potato and to help us grow this app to be available globally. For this we need global Agriculture experts to guide us select best features for our model and improve overall performance for offline usage. For the smart internet usage we need mentors from Google, together we can devise methods to achieve the goal while still conform to the Android restrictions.
Date | Task |
---|---|
December 28, 2019 | Our UI/UX has already been completed. |
January 15, 2020 | Dataset has already been collected from multiple sources and model is expected to be trained for offline use. |
January 30, 2020 | Expected testing and deployment in the TensorFlow lite model for initial beta. |
February 8, 2020 | To incorporate the Firebase ML Kit for working in online conditions and for the purpose of uploading samples that couldn’t be recognized by model, for future incremental learning. |
March 28, 2020 | To complete rest of online features. |
April 25, 2020 | Smart Internet Usage in working condition |
We are a team of three members. We are helping connect farmers and make use of AI for their productivity.
Myself Mirwise Khan, an enthusiastic Java developer, UI/UX designer and a freelancer. I’ve worked on diverse set of projects including, movies review sentiment analysis, tweets sentiment analysis, created REST APIs, JavaFX projects.