<![CDATA[This study presents a comparative analysis between the Support Vector Machine (SVM) algorithm, representing machine learning techniques, and the Convolutional Neural Network (CNN) algorithm, representing deep learning techniques, for vegetable image classification. The research adopts a quantitative approach by conducting multiple experimental schemes involving various SVM feature extraction methods and CNN architectures. The objective of this study is to evaluate and compare the performance, effectiveness, and computational efficiency of SVM and CNN algorithms. The dataset used in this study is a publicly available vegetable image dataset obtained from the Kaggle platform, consisting of 21.000 images categorized into 15 classes. The experimental results indicate that CNN significantly outperforms SVM in terms of accuracy, precision, recall, and F1-score. Moreover, CNN demonstrates superior generalization capability in predicting unseen image data. The best performance of the SVM algorithm was achieved using the Color Histogram feature extraction method, yielding an accuracy of 93%. In contrast, CNN models employing pre-trained architectures achieved higher accuracy, with VGG16 and MobileNetV2 obtaining accuracies of 98% and 100%, respectively. Based on the comparative results, CNN provides higher classification accuracy than SVM; therefore, this study can serve as a scientific reference for the development of image classification systems in digital agriculture and other applications requiring high accuracy and efficient computational performance.]]>
Copyrights © 2026
