About this document
House-GAN: Graph-Based Layout Generation by mas.subscrib is a document available to read on EtoBox.
The document presents House-GAN, a novel graph-constrained generative adversarial network designed for generating realistic house layouts based on input bubble diagrams. It utilizes a relational architecture to encode architectural constraints into the graph structure, allowing for the production of diverse and compatible house layouts. Evaluations demonstrate that House-GAN outperforms existing methods in generating realistic floorplans, and the authors plan to share their code and data publicly.
- Author
- mas.subscrib
- Language
- EN