Identification of Medicinal and Cosmetic Plant Using Deep Learning

Snehal H. Pimple, Dr. Ashwini S. Gaikwad

Abstract: Accurate identification of medicinal and cosmetic plants is essential in healthcare, traditional medicine, herbal research, and cosmetic manufacturing. Conventional identification methods depend on botanical expertise and manual observation, which may lead to misclassification when performed by non-specialists. This paper presents the detailed implementation of a deep learning-based plant identification system capable of classifying more than 50 medicinal and cosmetic plant species. The proposed system employs transfer learning using MobileNetV2 to balance computational efficiency and classification accuracy. The model achieved 87.8% validation accuracy with an average prediction time below three seconds. The implementation includes dataset preparation, preprocessing, augmentation, model training, fine-tuning, performance evaluation, and deployment using a web interface. The system also integrates safety information and dual-purpose categorization to enhance practical usability. Experimental results demonstrate that lightweight convolutional neural networks can effectively support botanical identification tasks for real-world applications.

Keywords: Deep Learning, Convolutional Neural Networks, Medicinal Plants, Cosmetic Plants, Transfer Learning, Image Classification

How to Cite?: Snehal H. Pimple, Dr. Ashwini S. Gaikwad, "Identification of Medicinal and Cosmetic Plant Using Deep Learning", Volume 15 Issue 5, May 2026, International Journal of Science and Research (IJSR), Pages: 1370-1373, https://www.ijsr.net/getabstract.php?paperid=SR26521163438, DOI: https://dx.dx.doi.org/10.21275/SR26521163438

Download Citation: APA | MLA | BibTeX | EndNote | RefMan