HierRelTriple: Guiding Indoor Layout Generation with Hierarchical Relationship Triplet Losses
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918141405167616 |
|---|---|
| author | Sun, Kaifan Yang, Bingchen Wonka, Peter Xiao, Jun Jiang, Haiyong |
| author_facet | Sun, Kaifan Yang, Bingchen Wonka, Peter Xiao, Jun Jiang, Haiyong |
| contents | We present a hierarchical triplet-based indoor relationship learning method, coined HierRelTriple, with a focus on spatial relationship learning. Existing approaches often depend on manually defined spatial rules or simplified pairwise representations, which fail to capture complex, multi-object relationships found in real scenarios and lead to overcrowded or physically implausible arrangements. We introduce HierRelTriple, a hierarchical relational triplets modeling framework that first partitions functional regions and then automatically extracts three levels of spatial relationships: object-to-region (O2R), object-to-object (O2O), and corner-to-corner (C2C). By representing these relationships as geometric triplets and employing approaches based on Delaunay Triangulation to establish spatial priors, we derive IoU loss between denoised and ground truth triplets and integrate them seamlessly into the diffusion denoising process. The introduction of the joint formulation of inter-object distances, angular orientations, and spatial relationships enhances the physical realism of the generated scenes. Extensive experiments on unconditional layout synthesis, floorplan-conditioned layout generation, and scene rearrangement demonstrate that HierRelTriple improves spatial-relation metrics by over 15% and substantially reduces collisions and boundary violations compared to state-of-the-art methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_20289 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | HierRelTriple: Guiding Indoor Layout Generation with Hierarchical Relationship Triplet Losses Sun, Kaifan Yang, Bingchen Wonka, Peter Xiao, Jun Jiang, Haiyong Computer Vision and Pattern Recognition We present a hierarchical triplet-based indoor relationship learning method, coined HierRelTriple, with a focus on spatial relationship learning. Existing approaches often depend on manually defined spatial rules or simplified pairwise representations, which fail to capture complex, multi-object relationships found in real scenarios and lead to overcrowded or physically implausible arrangements. We introduce HierRelTriple, a hierarchical relational triplets modeling framework that first partitions functional regions and then automatically extracts three levels of spatial relationships: object-to-region (O2R), object-to-object (O2O), and corner-to-corner (C2C). By representing these relationships as geometric triplets and employing approaches based on Delaunay Triangulation to establish spatial priors, we derive IoU loss between denoised and ground truth triplets and integrate them seamlessly into the diffusion denoising process. The introduction of the joint formulation of inter-object distances, angular orientations, and spatial relationships enhances the physical realism of the generated scenes. Extensive experiments on unconditional layout synthesis, floorplan-conditioned layout generation, and scene rearrangement demonstrate that HierRelTriple improves spatial-relation metrics by over 15% and substantially reduces collisions and boundary violations compared to state-of-the-art methods. |
| title | HierRelTriple: Guiding Indoor Layout Generation with Hierarchical Relationship Triplet Losses |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.20289 |