Determination of Medicinal Leaf Properties

This paper presents a state-of-the-art approach to determine the medicinal properties of leaves by employing advanced deep learning models, including Inception, ResNet, VGG, and EfficientNet, and developing a Flutter-based user interface. The study utilized two large-scale datasets, Leafsnap and Leafsnap30, comprising over 45,000 images of leaves, and associated medicinal properties, to train and evaluate the deep learning models’ performance. The system’s accuracy and efficacy were evaluated through extensive experimentation, demonstrating high precision and recall rates, with EfficientNet exhibiting the best performance. The Flutter-based user interface provides seamless interaction and search capabilities for medicinal properties based on image upload or