Clove Quality Classification using the ResNet50V2 Architecture

The development of artificial intelligence (AI) and computer vision has opened new opportunities in the agricultural industry, including in clove quality classification. The quality of cloves affects their market value and export standards; however, classification still faces challenges such as similarities in shape, size, and color, as well as variability in lighting and background images that can reduce the accuracy of classification models. To address these challenges, this study develops a clove quality classification system using a Convolutional Neural Network (CNN) with the ResNet50V2 architecture, which has proven to be superior in image processing. The dataset used consists of 1,250 images of cloves that were processed through stages of background removal, image cropping, and resizing to 224x224 pixels to meet the model's requirements. The data is divided into 80% for training and 20% for testing. The model is trained using deep learning techniques, with parameters optimized to enhance classification performance. The results show that the ResNet50V2 model achieves an accuracy of 98.80%, with very high precision, recall, and F1-score. The accuracy and loss graphs indicate that the model operates stably without experiencing overfitting, while the confusion matrix shows a very low prediction error rate. These results demonstrate that ResNet50V2 is effective in classifying clove quality.