We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Creative Painting with Latent Diffusion Models

Authors: Xianchao Wu
Abstract: Artistic painting has achieved significant progress during recent years. Using an autoencoder to connect the original images with compressed latent spaces and a cross attention enhanced U-Net as the backbone of diffusion, latent diffusion models (LDMs) have achieved stable and high fertility image generation. In this paper, we focus on enhancing the creative painting ability of current LDMs in two directions, textual condition extension and model retraining with Wikiart dataset. Through textual condition extension, users' input prompts are expanded with rich contextual knowledge for deeper understanding and explaining the prompts. Wikiart dataset contains 80K famous artworks drawn during recent 400 years by more than 1,000 famous artists in rich styles and genres. Through the retraining, we are able to ask these artists to draw novel and creative painting on modern topics. Direct comparisons with the original model show that the creativity and artistry are enriched.
Comments: 17pages, 12 figures
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Graphics (cs.GR); Machine Learning (cs.LG)
Cite as: arXiv:2209.14697 [cs.CV]
  (or arXiv:2209.14697v2 [cs.CV] for this version)

Submission history

From: Xianchao Wu [view email]
[v1] Thu, 29 Sep 2022 11:49:07 GMT (13676kb,D)
[v2] Fri, 30 Sep 2022 03:08:34 GMT (3367kb,D)

Link back to: arXiv, form interface, contact.