Semantic Relation Model and Dataset for Remote Sensing Scene Understanding

A deep understanding of our visual world is more than an isolated perception on a series objects, and the relationships between them also contain rich semantic information. Especially for those satellite remote sensing images, span so large that various objects are always different sizes complex spatial compositions. Therefore, recognition relations conducive to strengthen scenes. In this paper, we propose novel multi-scale fusion network (MSFN). framework, dilated convolution introduced into graph convolutional (GCN) based attentional mechanism fuse refine context, which crucial cognitive ability model Besides, mapping features embeddings, design sparse relationship extraction module remove meaningless connections among entities improve efficiency scene generation. Meanwhile, further promote research in field, paper proposes dataset (RSSGD). We carry out extensive experiments results show significantly outperforms previous methods addition, RSSGD effectively bridges huge gap low-level high-level cognition images.