Fig 1.
Interior design images generated by different models.
Fig 2.
Illustration of the proposed DiffDesign.
Fig 3.
Examples of DesignHelper dataset for public spaces.
Fig 4.
Examples of DesignHelper dataset for home and leisure spaces.
Fig 5.
Interior design images generated by our diffusion model for different decoration styles and space functions.
Fig 6.
Interior design images generated by different diffusion models for different space functions.
Table 1.
Zero-shot image-text retrieval results on Flickr and MSCOCO datasets.
The best results are marked in bold.
Table 2.
Zero-shot image-text retrieval results on DesignHelper datasets.
The best results are marked in bold.
Table 3.
Comparison of different models based on CLIP Sim and IS and FID across English dataset. The best results are marked in bold.
Fig 7.
Computational efficiency evaluation.
Fig 8.
Average scores of human preference for six interior design solutions generation models.
Fig 9.
Renderings of different models generated based on the same prompt.
Fig 10.
Ablation study of the three modules of the proposed method, i.e., Mclip, Mga, and Mds.
Table 4.
A brief critical analysis of DiffDesign.