AI-Based Medicinal Plant Identification And Therapeutic Recommendation System

The proposed AI-Based Medicinal Plant Identification and Therapeutic Recommendation System combines deep learning and web technologies to identify medicinal plants from leaf images. The system uses a dataset containing 80 categories of Indian medicinal plants. Three deep learning architectures, namely EfficientNet-B0, ResNet50, and a hybrid model, are evaluated for classification performance. According to the reported experimental results, EfficientNet-B0 achieved the highest test accuracy of 99.95%. The system processes uploaded images and predicts the corresponding plant category using the trained model. A Flask-based web application developed with HTML, CSS, and JavaScript allows users to access the identification results conveniently. The application also provides associated botanical and medicinal information from its knowledge base. This approach can support botanical education, preliminary plant identification, and research activities. However, system performance depends on dataset quality, image conditions, and the similarity between plant species. Future improvements may include expanding the dataset with additional medicinal plants and incorporating images of flowers, stems, and fruits. Developing mobile applications could improve accessibility in outdoor environments. Regional language support, including Kannada and Hindi, could help reach more users. Further research should evaluate the model using independently collected images under different environmental conditions. The plant information database can also be improved using reliable botanical and scientific references. Adding expert verification and warnings for uncertain predictions could help reduce identification risks. Overall, the proposed system provides a foundation for accessible medicinal plant identification, although further validation is necessary before relying on its predictions for practical or healthcare-related decisions.

[1] Tan, M., and Le, Q. V., “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” Proceedings of the 36th International Conference on Machine Learning (ICML), pp. 6105–6114, 2019. [2] He, K., Zhang, X., Ren, S., and Sun, J., “Deep Residual Learning for Image Recognition,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, 2016. [3] Krizhevsky, A., Sutskever, I., and Hinton, G. E., “ImageNet Classification with Deep Convolutional Neural Networks,” Advances in Neural Information Processing Systems, vol. 25, 2012. [4] Simonyan, K., and Zisserman, A., “Very Deep Convolutional Networks for Large-Scale Image Recognition,” International Conference on Learning Representations (ICLR), 2015. [5] Howard, A. G., et al., “MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications,” arXiv preprint arXiv:1704.04861, 2017. [6] Chollet, F., “Xception: Deep Learning with Depthwise Separable Convolutions,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1251–1258, 2017. [7] LeCun, Y., Bengio, Y., and Hinton, G., “Deep Learning,” Nature, vol. 521, pp. 436–444, 2015. [8] Goodfellow, I., Bengio, Y., and Courville, A., Deep Learning. MIT Press, 2016. [9] Shorten, C., and Khoshgoftaar, T. M., “A Survey on Image Data Augmentation for Deep Learning,” Journal of Big Data, vol. 6, article 60, 2019. [10] Szeliski, R., Computer Vision: Algorithms and Applications, 2nd ed. Springer, 2022.