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

The accurate identification and classification of medicinal plants are crucial for botanical research, pharmacology, and traditional medicine, where wrong identification or categorization of the plant species may lead to worse medical effects. In this research, MediFlora-Net, a novel Deep Learning (DL) model is created and ideal for the accurate identification of medicinal plants. The proposed MediFlora-Net uses multi-modal DL methodologies, quantum-assisted feature extraction and hybrid ensembling methodologies in constructing the plant recognition model. Besides, the methodology uses Vision Transformer (ViT), Convolutional Neural Networks (CNNs) and Proposed Med-Plant-Generative Adversarial Networks (GANs). This makes the framework to be capable of handling multiple imaging modalities such as RGB, and Hyperspectral Botanical Imagery. Also, a new quantum-inspired feature extraction technique is integrated into the model in which quantum probabilistic feature mapping and entanglement-based representation are utilized to extract higher-order botanical features. The framework also includes separate 'feature fusion', fine-tuned attention, and probabilistic decision-making. The proposed MediFlora-Net advances medicinal plant identification to greater precision and flexibility for practical use in the conservation of biological diversity, ethnobotanical studies, and pharmacology. This work effectively exploits DL techniques and quantum-inspired approaches to tackle the inherent issues of botanical identification to enable the design of better-advanced systems of plant identification. The implementation and source code are available at https://github.com/kvuma02-svg/MEDICINAL-PLANT-IDENTIFICATION.