Grapevine Augmentation and Classification using Enhanced EfficientNetB5 Model

All over the world, grapevine leaves available in environment are used in a broad variety of traditional food types. Recognizing different kinds of grapevine leaves is turning into a need because their cost and flavour vary from kind to kind. Also, it takes time and effort for a human to distinguish between distinct grapevine leaf kinds. So, it is very important to create a machine learning model to categorize the classification of grapevine. As a result, this is the main subject of this research. In order to categorise grape leaves, this work converts the 86% accurate EfficientNetB5 model to a CNN-based model. The grid search hyperparameter tuning method is used to get the best values for a inclusive machine learning model. A range of values for each hyperparameter are defined before the model and is trained with every possible combination of those values. The classification success rates for the EfficientNetB5 models were 86% for accuracy prediction. The outcomes show that the categorization models for grapevine kinds utilised in the study can be successfully implemented in these kind of research work for welfare of the society.