参考论文

Triplet-Aware Scene Graph Embeddings

论文链接

Introduction

场景图(Scene graphs)作为一种结构数据用于描述两两物体之间的语义关系。在Scene graph中,每个节点代表一个物体,每一条边代表两个物体之间的关系,用一个三元组<subject, predicate, object>来表示。例如:<cat, on, road>,<dog, left of, person>。对于场景图的解释如下图所示(D.Xu, Y.Zhu, C.Choy and L.Fei-Fei. Scene graph generation by iterative message passing. In Computer Vision and Pattern Recongnition, 2017):

Scene Graph