About

Plant Leaf Disease Detection is an application made to detect and identify plant leaf diseases and find a cure. The deep learning models were created using Tensorflow and Keras open-source libraries. First, the models were trained by using the leaf databases (found in Kaggle, FiveThirtyEight.com, Data.gov, data-is-plural.com, Quandl.com, Socrata.com and Google Public Datasets).
The models were continuously trained with as many datasets as we could find and a sharp increase in accuracy curve was observed. Simultaneously, the frontend was created using Flutter and some of most modern libraries available for Android and IOS devices. A Flask-Ngrok hybrid server was used to connect the Frontend and Backend together.
This enabled us to achieve state-of-the-art deep learning capabilities and reach unprecedented accuracies for a mobile based deep learning model. The frontend itself boasts a fully Material Design taken straight from the material.io library for flutter. This ensures that this frontend is easy to learn and use, while also providing great functionality and user experience.
This would be a significant step towards digitalization of the agricultural industries of India, which have remained stagnant in terms of digitalization. This is also a significant step considering the low production costs and we hope that this can be delivered free of cost to the poor farmers who suffer greatly from the diseases of plant leaves.
So far, we have 5 working models that detect diseases for 5 different leaves. More models are being trained and we hope to deploy them and continuously integrate with our frontend.