Medicinal Plant Classification Using Transfer Learning with ResNet50

Authors:

1. Y. Madhuri¹, Dr. B.Rama Ganesh²

¹P.G Scholar,Department of Computer Science and Engineering, Sri Venkatesa Perumal College of Engineering & Technology, Puttur, Andhra Pradesh, India, madhuri.yandra94@gmail.com

²Professor&HoD, Department of Computer Science & Engineering, Sri Venkatesa Perumal College of Engineering & Technology, Puttur, Andhra Pradesh, India, ramaganesh34@gmail.com

Abstract -Medicinal plants have long played a vital role in traditional healthcare systems, yet accurate identification remains challenging due to visual similarities, regional variations, and language barriers. This project introduces an intelligent system leveraging Convolution Neural Networks (CNNs) with Transfer Learning using ResNet50 for medicinal plant identification from images. By fine-tuning ResNet50 on a labelled dataset of medicinal plant images, the system achieves high classification accuracy while efficiently extracting discriminative features such as leaf veins and texture patterns. Once a plant is identified, the platform retrieves ethnopharmacological insights and displays its scientific name, regional name, medicinal uses, parts used, and preparation methods. Multilingual support in both English and Telugu ensures accessibility for diverse users, including researchers, herbal practitioners, and local communities. The system addresses challenges such as image variability, class imbalance, and real-time identification requirements through optimized deep learning models and an intuitive user interface. By integrating artificial intelligence with ethno botanical knowledge, this project contributes to preserving traditional medicinal wisdom while enhancing its accessibility through modern technology. The scalability of the platform ensures usability across different devices, making it a valuable tool for education, research, and practical applications in traditional medicine.

Key Words: Machine Learning, Deep Learning, Transfer Learning, Convolution Neural Network, Plant Disease identification, ResNet 50.