MediFlora-Net: Quantum-enhanced deep learning for precision medicinal plant identification

Area of Science:

• Botany
• Computer Science
• Pharmacology

Background:

• Accurate identification of medicinal plants is critical for research and traditional medicine.
• Misidentification can lead to adverse medical outcomes.
• Existing methods may lack precision and flexibility.

Purpose of the Study:

• To develop a novel Deep Learning (DL) model, MediFlora-Net, for accurate medicinal plant identification.
• To integrate multi-modal DL, quantum-assisted feature extraction, and hybrid ensembling.
• To enhance the precision and flexibility of plant recognition systems.

Main Methods:

• Developed MediFlora-Net using Vision Transformer (ViT), Convolutional Neural Networks (CNNs), and Generative Adversarial Networks (GANs).
• Employed quantum-inspired feature extraction with probabilistic mapping and entanglement-based representation.
• Utilized multi-modal imaging (RGB, Hyperspectral) and incorporated feature fusion, attention, and probabilistic decision-making.

Main Results:

• MediFlora-Net demonstrates high precision in identifying and classifying medicinal plants.
• The model effectively handles multiple imaging modalities.
• Quantum-inspired techniques enhance the extraction of complex botanical features.

Conclusions:

• MediFlora-Net advances medicinal plant identification accuracy and flexibility.
• The study highlights the potential of combining DL and quantum-inspired approaches for botanical identification.
• This work supports applications in biodiversity conservation, ethnobotany, and pharmacology.