Computational Intelligence and Machine Learning
AI-Driven Recognition of Indian Medicinal Flora using Convolutional Neural Networks
Authors: Rajlaxmi Roy, Rajrupa Roy and Debasmita Chatterjee
Publishing Date: 23-04-2025
ISBN: 978-81-975670-5-6
Abstract
Plants, people, and culture have always been deeply intertwined, each influencing the other in maintaining an equilibrium that has shaped centuries of traditional healing practices. Identifying plants with healing properties requires extensive knowledge and experience, and relies heavily on human perception, which can lead to errors. Leveraging Deep Learning with Convolutional Neural Networks and transfer learning to overcome these limitations, we have worked to develop an integrated system for the identification of medicinal plants native to India. Our system identifies 98 different classes using a dataset of 15,022 real-world images of weeds, herbs, shrubs, and trees with 92% accuracy and gives the therapeutic uses of each identified plant. This approach will empower people with the knowledge necessary to preserve and enhance traditional medicinal practices, bridging the gap between age-old practice and integrative modern medicine. This system uses Xception as the base model with a custom classification layer which can be modified to ensure scalability, by increasing the dataset to include medicinal plant species found locally and their medicinal benefits. Further development of this system into an application will ensure comprehensiveness of usage and a broader impact on medicinal practices and biodiversity conservation initiatives.
Keywords
Computer Vision, Convolutional Neural Networks, Machine Learning, Medicinal Plants of India
Cite as
Rajlaxmi Roy, Rajrupa Roy and Debasmita Chatterjee, "AI-Driven Recognition of Indian Medicinal Flora using Convolutional Neural Networks", In: Sandeep Kumar and Kavita Sharma (eds), Computational Intelligence and Machine Learning, SCRS, India, 2025, pp. 21-33. https://doi.org/10.56155/978-81-975670-5-6-3