Automated Recognition of Medicinal Plants in the Wild: A Leaf-centric Approach

Abstract

This study explores the use of technology to simplify the identification of medicinal plants in the wild by focusing on leaf characteristics. Using convolutional neural networks (CNNs), the research aims to develop a mobile-friendly system tailored to Malaysia’s rich biodiversity and traditional medicine heritage. Key steps include collecting a diverse range of plant data, enhancing image quality through pre-processing, and testing various CNN models to determine the most effective one. Designed for use by both experts and non-experts, such as rural communities and herbalists, the tool integrates advanced AI with traditional knowledge to preserve cultural practices, promote safe natural remedies, and raise awareness about medicinal plants’ role in healthcare and conservation. By addressing the decline in herbal knowledge, this project aims to deliver a practical and accessible solution that supports public health and environmental sustainability.