<![CDATA[Salad is a type of leafy vegetable belonging to the Compositae family, genus Lactuca. It is rich in nutrients, including fiber, vitamin A, and minerals. Salad greens can cleanse the blood and fat, help people with coughs, and prevent high cholesterol, constipation, and insomnia. With the increasing population and awareness of the benefits of a balanced diet, consumer demand for lettuce has significantly increased. Consequently, farmers have expanded lettuce cultivation to meet consumer demand, which has the potential to cause errors in sorting different types of lettuce. Therefore, research focusing on the classification and detection of lettuce varieties is crucial to help farmers efficiently harvest lettuce based on its type using Convolutional Neural Network (CNN) methods, comparing three models: VGG16, Xception, and NasNetMobile. Data were directly obtained from lettuce farms and Kaggle. After pre-processing steps such as resizing and augmentation, the data were trained with various amounts, 200 epochs, and 64 batches during the architectural modeling stage. Based on the research results, the accuracy analysis with image classification of various types of lettuce concluded that using the Convolutional Neural Network (CNN) method by comparing three models VGG16, Xception, and NasNetMobile can classify each type of lettuce based on its class with high accuracy. In the tests conducted on the trained model, using an input size of 120 x 120, 200 epochs, and a batch size of 64, the NasNetMobile architecture model achieved the highest scores with an accuracy of 98.33%, precision of 97.8%, recall of 97.9%, and an F1-score of 97.8%. With these excellent accuracy values, the researchers hope that this analysis will make a significant contribution to the development of a superior and more efficient image classification system for agriculture, especially in selecting the appropriate CNN architecture.]]>
