MatFuse: Controllable Material Generation with Diffusion Models
This work simplifies material creation for graphics professionals, offering fine-grained control and editing capabilities, though it is incremental as it builds on existing diffusion models.
The paper tackles the challenge of creating high-quality 3D materials in computer graphics by introducing MatFuse, a diffusion model-based method that generates and edits materials using multiple conditioning inputs like color palettes and text, achieving improved CLIP-IQA and FID scores and positive user study results.
Creating high-quality materials in computer graphics is a challenging and time-consuming task, which requires great expertise. To simplify this process, we introduce MatFuse, a unified approach that harnesses the generative power of diffusion models for creation and editing of 3D materials. Our method integrates multiple sources of conditioning, including color palettes, sketches, text, and pictures, enhancing creative possibilities and granting fine-grained control over material synthesis. Additionally, MatFuse enables map-level material editing capabilities through latent manipulation by means of a multi-encoder compression model which learns a disentangled latent representation for each map. We demonstrate the effectiveness of MatFuse under multiple conditioning settings and explore the potential of material editing. Finally, we assess the quality of the generated materials both quantitatively in terms of CLIP-IQA and FID scores and qualitatively by conducting a user study. Source code for training MatFuse and supplemental materials are publicly available at https://gvecchio.com/matfuse.