Search papers, labs, and topics across Lattice.
This paper introduces Hyperbolic Scene Graph (HSG), a novel approach to learning scene graph embeddings by leveraging hyperbolic space to explicitly capture hierarchical relationships between objects and places. HSG uses geometric distance in hyperbolic space to encode these relationships, improving the structural consistency of the learned representations. Experiments demonstrate that HSG significantly outperforms Euclidean-based methods, particularly in graph-level metrics, achieving a PP IoU of 33.17 and a Graph IoU of 33.51, a substantial improvement over the best baseline.
Encoding scene graphs in hyperbolic space unlocks significantly better hierarchical structure quality compared to Euclidean embeddings, leading to state-of-the-art graph-level scene understanding.
Scene graph representations enable structured visual understanding by modeling objects and their relationships, and have been widely used for multiview and 3D scene reasoning. Existing methods such as MSG learn scene graph embeddings in Euclidean space using contrastive learning and attention based association. However, Euclidean geometry does not explicitly capture hierarchical entailment relationships between places and objects, limiting the structural consistency of learned representations. To address this, we propose Hyperbolic Scene Graph (HSG), which learns scene graph embeddings in hyperbolic space where hierarchical relationships are naturally encoded through geometric distance. Our results show that HSG improves hierarchical structure quality while maintaining strong retrieval performance. The largest gains are observed in graph level metrics: HSG achieves a PP IoU of 33.17 and the highest Graph IoU of 33.51, outperforming the best AoMSG variant (25.37) by 8.14, highlighting the effectiveness of hyperbolic representation learning for scene graph modeling. Code: https://github.com/AIGeeksGroup/HSG.
