SceneDiffusion: Scene Generation Model Embedded with Spatial Constraints
Spatial scenes, as fundamental units of geospatial cognition, encompass rich objects and spatial relationships, and their generation techniques hold significant application value in disaster simulation and emergency drills, delayed spatial reconstruction and analysis, and other fields. However, exis...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-06-01
|
| Series: | ISPRS International Journal of Geo-Information |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2220-9964/14/7/250 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849418141999824896 |
|---|---|
| author | Shanshan Yu Jiaxin Zhu Jiaqi Li Xunchun Li Kai Wang Jian Tu Danhuai Guo |
| author_facet | Shanshan Yu Jiaxin Zhu Jiaqi Li Xunchun Li Kai Wang Jian Tu Danhuai Guo |
| author_sort | Shanshan Yu |
| collection | DOAJ |
| description | Spatial scenes, as fundamental units of geospatial cognition, encompass rich objects and spatial relationships, and their generation techniques hold significant application value in disaster simulation and emergency drills, delayed spatial reconstruction and analysis, and other fields. However, existing studies still face limitations in modeling complex spatial relationships during scene generation, leading to insufficient semantic consistency and geographical accuracy. The advancement of Geospatial Artificial Intelligence (GeoAI) offers a new technical pathway for the intelligent modeling of spatial scenes. Against this backdrop, we propose SceneDiffusion, a scene generation model embedded with spatial constraints, and construct a geospatial scene dataset incorporating spatial relationship descriptions and geographic semantics, aiming to enhance the understanding and modeling capabilities of GeoAI models for spatial information. Specifically, SceneDiffusion employs a spatial scene representation framework to uniformly characterize objects and their topological, directional, and distance relationships, enhances the interactive modeling of objects and relationships through a Spatial relationship Attention-aware Graph (SAG) module, and finally generates high-quality scene images conforming to geographic semantics using a Layout information-guided Conditional Diffusion (LCD) module. Both qualitative and quantitative experiments demonstrate the superiority of SceneDiffusion, achieving a 56.6% reduction in FID and a 35.3% improvement in SSIM compared to baseline methods. Ablation studies confirm the importance of multi-relational modeling with attention mechanisms. By generating scenes that satisfy spatial distribution constraints, this work provides technical support for applications such as emergency scene simulation and virtual scene construction, while also offering insights for theoretical research and methodological innovation in GeoAI. |
| format | Article |
| id | doaj-art-73cc6d3a0c4b4e0d87121d6243edba92 |
| institution | Kabale University |
| issn | 2220-9964 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | ISPRS International Journal of Geo-Information |
| spelling | doaj-art-73cc6d3a0c4b4e0d87121d6243edba922025-08-20T03:32:32ZengMDPI AGISPRS International Journal of Geo-Information2220-99642025-06-0114725010.3390/ijgi14070250SceneDiffusion: Scene Generation Model Embedded with Spatial ConstraintsShanshan Yu0Jiaxin Zhu1Jiaqi Li2Xunchun Li3Kai Wang4Jian Tu5Danhuai Guo6School of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, ChinaSchool of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, ChinaSchool of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, ChinaAcademy of Broadcasting Science, National Radio and Television Administration of China, Beijing 100866, ChinaBeijing Institute of Aerospace Long March Vehicles, Beijing 100076, ChinaBeijing Institute of Aerospace Long March Vehicles, Beijing 100076, ChinaSchool of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, ChinaSpatial scenes, as fundamental units of geospatial cognition, encompass rich objects and spatial relationships, and their generation techniques hold significant application value in disaster simulation and emergency drills, delayed spatial reconstruction and analysis, and other fields. However, existing studies still face limitations in modeling complex spatial relationships during scene generation, leading to insufficient semantic consistency and geographical accuracy. The advancement of Geospatial Artificial Intelligence (GeoAI) offers a new technical pathway for the intelligent modeling of spatial scenes. Against this backdrop, we propose SceneDiffusion, a scene generation model embedded with spatial constraints, and construct a geospatial scene dataset incorporating spatial relationship descriptions and geographic semantics, aiming to enhance the understanding and modeling capabilities of GeoAI models for spatial information. Specifically, SceneDiffusion employs a spatial scene representation framework to uniformly characterize objects and their topological, directional, and distance relationships, enhances the interactive modeling of objects and relationships through a Spatial relationship Attention-aware Graph (SAG) module, and finally generates high-quality scene images conforming to geographic semantics using a Layout information-guided Conditional Diffusion (LCD) module. Both qualitative and quantitative experiments demonstrate the superiority of SceneDiffusion, achieving a 56.6% reduction in FID and a 35.3% improvement in SSIM compared to baseline methods. Ablation studies confirm the importance of multi-relational modeling with attention mechanisms. By generating scenes that satisfy spatial distribution constraints, this work provides technical support for applications such as emergency scene simulation and virtual scene construction, while also offering insights for theoretical research and methodological innovation in GeoAI.https://www.mdpi.com/2220-9964/14/7/250GeoAIscene generationspatial constraintsdiffusion modelsspatial relationship modelingattention mechanisms |
| spellingShingle | Shanshan Yu Jiaxin Zhu Jiaqi Li Xunchun Li Kai Wang Jian Tu Danhuai Guo SceneDiffusion: Scene Generation Model Embedded with Spatial Constraints ISPRS International Journal of Geo-Information GeoAI scene generation spatial constraints diffusion models spatial relationship modeling attention mechanisms |
| title | SceneDiffusion: Scene Generation Model Embedded with Spatial Constraints |
| title_full | SceneDiffusion: Scene Generation Model Embedded with Spatial Constraints |
| title_fullStr | SceneDiffusion: Scene Generation Model Embedded with Spatial Constraints |
| title_full_unstemmed | SceneDiffusion: Scene Generation Model Embedded with Spatial Constraints |
| title_short | SceneDiffusion: Scene Generation Model Embedded with Spatial Constraints |
| title_sort | scenediffusion scene generation model embedded with spatial constraints |
| topic | GeoAI scene generation spatial constraints diffusion models spatial relationship modeling attention mechanisms |
| url | https://www.mdpi.com/2220-9964/14/7/250 |
| work_keys_str_mv | AT shanshanyu scenediffusionscenegenerationmodelembeddedwithspatialconstraints AT jiaxinzhu scenediffusionscenegenerationmodelembeddedwithspatialconstraints AT jiaqili scenediffusionscenegenerationmodelembeddedwithspatialconstraints AT xunchunli scenediffusionscenegenerationmodelembeddedwithspatialconstraints AT kaiwang scenediffusionscenegenerationmodelembeddedwithspatialconstraints AT jiantu scenediffusionscenegenerationmodelembeddedwithspatialconstraints AT danhuaiguo scenediffusionscenegenerationmodelembeddedwithspatialconstraints |