Semantic scene segmentation is a critical component of robot autonomy, enabling robots to understand and interact with their environment. This chapter introduces semantic segmentation, a dense prediction task that assigns a semantic class label to each pixel in an image, providing a detailed scene representation. It discusses various deep learning techniques, algorithms, and architectures used for semantic segmentation, along with loss functions, datasets, and benchmarks. The chapter also covers challenges and opportunities in this area, including the importance of semantic information for robotics applications such as autonomous driving, robot-assisted surgery, and search and rescue. It explores different perception tasks, including semantic segmentation, instance segmentation, and panoptic segmentation, and their respective roles in scene understanding. The chapter further discusses techniques for exploiting context, such as image pyramids, spatial pyramid pooling, and dilated convolutions, which enhance the performance of semantic segmentation. It also addresses real-time architectures and multimodal semantic segmentation, which combine data from multiple sources to improve robustness and accuracy. Finally, the chapter reviews public datasets and benchmarks for semantic segmentation, highlighting their importance in advancing the field.Semantic scene segmentation is a critical component of robot autonomy, enabling robots to understand and interact with their environment. This chapter introduces semantic segmentation, a dense prediction task that assigns a semantic class label to each pixel in an image, providing a detailed scene representation. It discusses various deep learning techniques, algorithms, and architectures used for semantic segmentation, along with loss functions, datasets, and benchmarks. The chapter also covers challenges and opportunities in this area, including the importance of semantic information for robotics applications such as autonomous driving, robot-assisted surgery, and search and rescue. It explores different perception tasks, including semantic segmentation, instance segmentation, and panoptic segmentation, and their respective roles in scene understanding. The chapter further discusses techniques for exploiting context, such as image pyramids, spatial pyramid pooling, and dilated convolutions, which enhance the performance of semantic segmentation. It also addresses real-time architectures and multimodal semantic segmentation, which combine data from multiple sources to improve robustness and accuracy. Finally, the chapter reviews public datasets and benchmarks for semantic segmentation, highlighting their importance in advancing the field.