A Customized Robust Deep Learning Approach for Efficient Medicinal Plant Recognition

DOI

10.2991/978-94-6239-664-7_55How to use a DOI?

Keywords

Medicinal plants; Deep learning; Transfer learning; Image classification; EfficientNetV2 B1

Abstract

Medicinal plants are a significant source of healthcare in both traditional and modern medicine, and their identification remains a significant challenge due to morphological similarities between foliage and the lack of comprehensive databases. Traditional classification methods, such as Support Vector Machines (SVM) and K-Nearest Neighbors (KNN), do not perform well under these conditions. To overcome these deficiencies, the current research paper presents a powerful and customized transfer learning model based on the EfficientNetV2_B1 backbone. It is a Squeeze-and-Excitation (SE) attention architecture, which scaled the architecture to improve the way features are represented and is trained together with 30 percent dropout and L2 weight regularization, which discourages overfitting. This model was trained and tested on three different sets of Bangladeshi medicinal plants. Data augmentation was performed to ensure the model is not readily influenced by the variations in position, size, brightness, and angle. The proposed individualized model demonstrated superior functioning compared to the baseline and the current practices and had the capacity to tackle the issues related to foliage similarity. The proposed model achieves 99.57%, 97.52%, and 99.60% accuracy on three distinct datasets. In addition, the model has been employed to achieve a macro and weighted F1 score of around one point, which is a measure of its reliability and accuracy in the classification of medicinal plants.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Ashari Binte Ashraf
AU  - Rakibul Haque Rabbi
AU  - Bishal Biswas
AU  - Shah Md Tanvir Siddiquee
PY  - 2026
DA  - 2026/06/08
TI  - A Customized Robust Deep Learning Approach for Efficient Medicinal Plant Recognition
BT  - Proceedings of the International Conference on Intelligent Data Analysis and Applications (IDAA 2025)
PB  - Atlantis Press
SP  - 799
EP  - 812
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-664-7_55
DO  - 10.2991/978-94-6239-664-7_55
ID  - Ashraf2026
ER  -