Medicinal Herb Identification through Stacking Ensemble Learning and Fine-Tuned Deep Neural Networks for Feature Learning

Abstract

Herbal medicine, as a primary healthcare system in developing countries, has garnered renewed attention owing to the valuable properties exhibited by herbs. Consequently, the classification of herbal images has emerged as a crucial research area. In this study, our aim is to enhance the classification of herbal medicine. Our proposed model harnesses the power of deep neural networks to train and automatically extract features from a comprehensive dataset of herbal images. Subsequently, a stacking ensemble technique is employed to classify these extracted features. Through numerical tests conducted on four distinct datasets of herbal medicinal plants, we demonstrate the efficacy of our proposed models. Notably, our models achieve improved accuracy compared to the studies referenced in this research, signifying the potential of our approach to advance the field of herbal medicine classification.