Generation of realistic floorplans using diffusion-based models

Thumbnail Image

Files

master_Hahkio_Linh_2023.pdf (9.29 MB)

URL

Journal Title

Journal ISSN

Volume Title

Perustieteiden korkeakoulu | Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.

Authors

Date

2023-10-09

Department

Major/Subject

Machine Learning, Data Science and Artificial Intelligence

Mcode

SCI3044

Degree programme

Master’s Programme in Computer, Communication and Information Sciences

Language

en

Pages

32

Series

Abstract

Existing studies on automatic floor plan generation have been mainly focused on the room layout of the floorplans while ignoring fenestration and furniture details, which was justified by the fact that these details are often irrelevant for the user. In this thesis, we propose to generate floorplans using diffusion-based generative models trained on images that contain fenestration and furniture information. Our experiments suggest this approach results in more realistic floorplans in comparison to previous generative models such as HouseGAN++. We also show that the proposed diffusion-based generative models can be used for reconstructing missing information in existing floorplans, for example, missing labels of the room types.

Description

Supervisor

Ilin, Alexander

Thesis advisor

Naderi, Kourosh

Keywords

floorplan generation, computer vision, generative model, diffusion model

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202310156363

Collections

[dipl] Perustieteiden korkeakoulu / SCI