Abstract: Semantic segmentation is a basic task in computer vision, especially in applications such as autonomous driving and scene understanding. While Transformer-Based models perform high-level ...