Automated Medicinal Plant Identification Using Deep Learning for Improved Healthcare

Abstract

With thousands of plant species present in the world, many of them possessing medicinal properties, the identification of these plants plays a crucial role in healthcare, drug manufacturing, and environmental management. Proper identification benefits various sectors, including the forest department, life scientists, physicians, medication laboratories, and the public. While manual identification by skilled practitioners has been the traditional method, it is often time-consuming and prone to misidentification, which could lead to side effects or serious health risks. To address these challenges, this project proposes an automated system for medicinal plant classification using deep learning algorithms. In recent years, the use of deep learning techniques, particularly in computer vision, has gained traction for solving complex identification problems. The proposed system employs convolutional neural networks (CNNs), which are highly effective in feature extraction and classification tasks. The model utilizes Xception-based feature extraction, followed by classification with a CNN classifier, demonstrating high accuracy and faster prediction times with real-time images. By automating the identification process, the system aims to enable faster and more reliable recognition of medicinal plant species. This advancement could significantly reduce errors, increase efficiency, and provide immediate access to valuable plant-based medicinal data for various stakeholders, ensuring safer and more effective use of medicinal plants in healthcare.

Introduction

Plants are multicellular organisms that produce food through photosynthesis and play a crucial role in ecosystems by generating oxygen and forming the base of many food chains. Medicinal plants contain substances used for therapeutic purposes and have been utilized for thousands of years across various cultures and traditional medicine systems like Unani, Ayurveda, and Chinese medicine.

Traditional plant identification, especially based on leaf morphology, is manual, time-consuming, and requires expert knowledge. This process involves field observation, use of botanical keys, morphological analysis, herbarium specimens, and dichotomous keys, but faces challenges like limited experts, long processing times, and difficulty distinguishing similar species.

The proposed system automates medicinal plant identification using leaf images through advanced image processing and machine learning, particularly convolutional neural networks (CNNs) and Xception features, enabling real-time, accurate classification. It also integrates a usage recommendation module to provide medicinal and practical insights about the identified plants. The system includes a user-friendly web app built with Python, Flask, MySQL, and Bootstrap, featuring secure user and admin interfaces, dataset management, continuous model training, and user plant identification with recommendations.

The LeafNet model uses publicly available leaf image datasets (like the Swedish dataset) and involves image preprocessing (grayscale conversion, noise reduction) for improved classification accuracy. Users upload leaf images, which the model classifies, and then receive detailed usage recommendations.

Testing showed the system is reliable, accurate, and efficient, reducing dependence on manual identification while promoting education and safe medicinal plant usage.

Conclusion

Classifying the plants species with their leaf using the algorithms of computer vision show concern on categorizing the plant images into its distinct groups. The classification of plants using digital leaf images are challenging due to their similarities in inter-class and intra-class, the possibility of complex background and variations in many parameters such as illumination and color. Thus, developing tools and solutions to analyze and interpret the patterns in the leaf images with significant results are essential. This project proposes an automated plant identification system, for identifying the plants species through their leaf. This task is accomplished using deep convolutional neural network to achieve higher accuracy. Image pre-processing, feature extraction and recognition are three main identification steps which are taken under consideration. Proposed CNN classifier learns the features of plants such as classification of leafs by using hidden layers like convolutional layer, max pooling layer, dropout layers and fully connected layers. The model acquires a knowledge related to features of Swedish leaf dataset in which 30 plant classes are available, that helps to predict the correct category of unknown plant with accuracy of 97% and minimum losses. Result is slightly better than the previous work that analyzes 93.75% of accuracy.

References

[1] S. Mahajan, A. Raina, X.-Z. Gao, and A. K. Pandit, ``Plant recognition using morphological feature extraction and transfer learning overSVMand AdaBoost,\'\' Symmetry, vol. 13, no. 2, p. 356, Feb. 2021. [2] Beikmohammadi, K. Faez, and A.Motallebi ``SWP-leaf NET: A novel multistage approach for plant leaf identification based on deep learning,\'\' 2020, arXiv:2009.05139. [Online]. Available: http://arxiv.org/abs/2009.05139. [3] Muneer and S. M. Fati, ``Efficient and automated herbs classification approach based on shape and texture features using deep learning,\'\' IEEE Access, vol. 8, pp. 196747-196764, 2020. [4] J. W. Lee and Y. C. Yoon, ``Fine-grained plant identification using wide and deep learning model 1,\'\' in Proc. Int. Conf. Platform Technol. Service (PlatCon), Jan. 2019, pp. 1-5. [5] P. Jasitha, M. R. Dileep, and M. Divya, ``Venation based plant leaves classification using GoogLeNet and VGG,\'\' in Proc. 4th Int. Conf. Recent Trends Electron., Inf., Commun. Technol. (RTEICT), May 2019, pp. 715-719.

Copyright

Copyright © 2025 Janakiraman S, Nandhini S. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.