Medical Plant Identification and Utilities Recommendation using Ensemble Techniques

This paper presents a deep learning-based system capable of classifying medicinal and Ayurvedic plants with high accuracy, countering the increasing decoupling with traditional knowledge as a result of increasing urbanization. Unlike traditional identification methods, which are generally manual and expert-dependent, the solution presented here provides high-accuracy classification as well as informative utility. The plant is not just identified, but other details like its scientific name, medicinal applications, and suggested home remedies are also provided. The system is implemented using a hybrid ensemble of InceptionV3 and ConvNeXtTiny models, trained on a heterogeneous data set of 30 medicinal plants in order to provide secure and reliable classification.To recommend appropriate treatments, a collaborative filtering approach is used to compare the identified plant to Ayurvedic treatments in a structured CSV database. The system is accessed via a Streamlit-based web portal and an Android smartphone app, enabling real-time, on-the-spot plant identification and remedy recommendation. Model performance is measured in terms of accuracy, which decides its deployment readiness. With the integration of traditional Ayurvedic practices and modern artificial intelligence, this system provides a pragmatic, user-friendly tool for promoting the use of natural medicine and public awareness of plant-based health remedies.