SSMA is a state-of-the-art deep learning fusion scheme for self-supervised multimodal semantic image segmentation, where the goal is to exploit complementary features from different modalities and ...