Abstract:
Plants, in all their diversity, represent a valuable contribution from the environment. When
illness strikes, whether physical or mental, medicine becomes our primary recourse. Many
pharmaceuticals are derived from natural plants, which serve as vital sources of medicinal
compounds. In Bangladesh, these medicinal plants are also known by the names
Homeopathy, Unani, and Ayurveda. Experts suggest that the COVID-19 pandemic can be
effectively addressed with the aid of medicinal plants. Strengthening our immune system
is paramount, as it directly influences overall health. A robust immune system can combat
bacteria, viruses, and other pathogens. Conversely, inactive individuals are more
susceptible to viral infections and diseases. Certain medicinal plants have been shown to
enhance immunity. Therefore, accurately classifying these plants is crucial.
We proposed using four well-known algorithms—DenseNet201, VGG19, and
ResNet152—to classify medicinal plants based on leaf images. Our approach achieved
accuracies of 96.08% with DenseNet201, 97% with VGG19, 98% with ResNet152, and an
impressive 99.12% with a hybrid model combining VGG19 and ResNet50. These results
highlight the potential of advanced deep learning techniques in enhancing the identification
and utilization of medicinal plants. Our research will help people understand their immune
systems better and the medicinal plants that can boost immunity, thereby enabling them to
combat diseases and viruses more effectively in the future. As we look ahead, it is
important to acknowledge that dangerous infectious viruses may continue to emerge. Our
study not only contributes to the field of medicinal plant classification but also underscores
the importance of integrating traditional medicinal knowledge with modern technological
advancements to improve public health outcomes.
