GLTScene: Global‐to‐Local Transformers for Indoor Scene Synthesis with General Room Boundaries
- 24 October 2024
- journal article
- research article
- Published by Wiley in Computer Graphics Forum
- Vol. 43 (7)
Abstract
We present GLTScene, a novel data‐driven method for high‐quality furniture layout synthesis with general room boundaries as conditions. This task is challenging since the existing indoor scene datasets do not cover the variety of general room boundaries. We incorporate the interior design principles with learning techniques and adopt a global‐to‐local strategy for this task. Globally, we learn the placement of furniture objects from the datasets without considering their alignment. Locally, we learn the alignment of furniture objects relative to their nearest walls, according to the alignment principle in interior design. The global placement and local alignment of furniture objects are achieved by two transformers respectively. We compare our method with several baselines in the task of furniture layout synthesis with general room boundaries as conditions. Our method outperforms these baselines both quantitatively and qualitatively. We also demonstrate that our method can achieve other conditional layout synthesis tasks, including object‐level conditional generation and attribute‐level conditional generation. The code is publicly available at https://github.com/WWalter-Lee/GLTScene.Keywords
Funding Information
- National Natural Science Foundation of China (62072316)
- Natural Science Foundation of Guangdong Province (2023A1515011297)
This publication has 14 references indexed in Scilit:
- DiffuScene: Denoising Diffusion Models for Generative Indoor Scene Synthesis2024
- MIME: Human-Aware 3D Scene Generation2023
- LEGO-Net: Learning Regular Rearrangements of Objects in Rooms2023
- SceneHGN: Hierarchical Graph Networks for 3D Indoor Scene Generation With Fine-Grained GeometryIEEE Transactions on Pattern Analysis and Machine Intelligence, 2023
- 3D-FUTURE: 3D Furniture Shape with TextUREInternational Journal of Computer Vision, 2021
- 3D-FRONT: 3D Furnished Rooms with layOuts and semaNTics2021
- GRAINSACM Transactions on Graphics, 2019
- Deep Residual Learning for Image Recognition2016
- Interactive furniture layout using interior design guidelinesACM Transactions on Graphics, 2011
- Make it homeACM Transactions on Graphics, 2011