Research Projects

The project aims to develop lightweight Deep Neural Network-based models for predicting plant diseases by analyzing input plant leaf images. The project will have a multilingual UI interface, allowing users to capture and send plant images to server side web API components. These components will process user request data in a transactional database and call the core image processing algorithm, which will analyze the image and predict plant diseases with high accuracy. The disease prediction will then be passed through a remediation module, creating an accurate, cost-effective disease treatment plan. The objectives of the project include developing an effective algorithm capable of predicting plant leaf disease using latest Deep Learning techniques, improving state-of-the-art research on plant leaf disease detection, and developing algorithms for disease modeling to artificially synthesize a particular disease or multiple diseases on a leaf image. The synthetic data set and trained models will be made publicly available for further research and development. The project will implement the algorithms/models developed through the project on an android platform and distribute it as an android-based smartphone application via APIs. The models will be optimized by reducing network parameters to make the software light-weight and allowing users to download and execute it even on devices with less computational power. Server side components will be created on private/public cloud infrastructure with web-services to connect/execute image processing deep learning algorithms and a transactional database to store relevant data for future use. The target beneficiaries include farmers, scientists, engineers, and Indian technology companies working in computer vision engineering. This could lead to future products based on image analysis, such as space exploration, marine engineering, and facial recognition. Collaborating institutions, agencies, and industries is encouraged.