Lettuce (Lactuca sativa L.) is a widely consumed vegetable known for its metabolic benefits and role in preventing chronic diseases. Due to the morphological complexity and variability in leaf characteristics among different cultivars, segmenting lettuce is essential for automated growth monitoring, early disease detection, and the optimization of agricultural practices. The presented paper aims to evaluate the effectiveness of classical Digital Image Processing (DIP) techniques in comparison to the YOLOv11n model for lettuce segmentation. For the analysis, images of Milena curly lettuce cultivars grown in a hydroponic system were captured. The segmentation process for DIP was carried out using multiple thresholding in the HSV color space. Meanwhile, the YOLOv11n model was trained in the Google Colaboratory environment using an NVIDIA A100 GPU. The performance of both approaches was assessed through metrics such as Intersection over Union (IoU), precision, recall, and execution time. The results revealed that the DIP techniques achieved average IoU, precision, and recall values of 0.81, 0.90, and 0.89, respectively. In contrast, the YOLOv11n model demonstrated superior performance, with IoU, precision, and recall values of 0.92, 0.95, and 0.96, respectively. However, DIP excelled in terms of efficiency, segmenting 10 images in just 0.32 s, compared to the 28.56 s required by YOLOv11n.