Grapevine Leaves Image Dataset

The first approach used a pre-trained [CNN model name], specifically fine-tuned for the classification task. In the second approach, features were extracted from the [CNN model name] model’s Logits layer, and classification was performed using multiple SVM kernel functions. The third approach implemented feature selection with the [feature selection method], reducing the feature set from [original number] to [reduced number]. Classification was then performed using the selected features with various SVM kernels.

The highest classification accuracy was achieved through feature extraction from the Logits layer and feature reduction via [feature selection method]. The best-performing SVM kernel was [kernel type]. The system achieved an overall classification accuracy of [accuracy]% using the [CNN model name]-SVM [kernel type] approach, demonstrating the importance of feature selection for improved classification accuracy.

This dataset is sourced from Kaggle.