Skip to main navigation Skip to search Skip to main content

Divide and control: generation of multiple component comic illustrations with diffusion models based on regression

  • Zixuan Wang
  • , Peng Du
  • , Zhenghui Xu
  • , Qihan Hu
  • , Hao Zeng
  • , Youbing Zhao
  • , Hao Xie
  • , Tongqing Ma
  • , Shengyou Lin
  • Communication University of Zhejiang
  • Uber Technologies, Inc.

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Diffusion-based text-to-image generation has achieved huge success in creative image generation and editing applications. However, when applied to comic illustrations, it still struggles to deliver predictable high-quality productions with multiple characters due to the interference of the text prompts. In this paper, we propose a practicable method to use ControlNet and stable diffusion to generate controllable outputs of multiple components. The method first generates images for individual components separately and then degenerates those images to a regressed form, such as line drawings or Canny edges. Those regressed forms of individual components are then merged and fed into ControlNet to generate the final image. Experiments show that this method is highly controllable and can produce high-quality comic illustrations with multiple components.
Original languageEnglish
Title of host publicationAI-generated Content - 1st International Conference, AIGC 2023, Revised Selected Papers
EditorsFeng Zhao, Duoqian Miao
PublisherSpringer
Pages59-69
Number of pages11
Edition1946
ISBN (Print)9789819975860
DOIs
Publication statusPublished - 2 Nov 2023
EventInternational Conference on AI-generated Content, AIGC 2023 - Shanghai
Duration: 25 Aug 202326 Aug 2023

Publication series

NameCommunications in Computer and Information Science
Volume1946 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

ConferenceInternational Conference on AI-generated Content, AIGC 2023
CityShanghai
Period25/08/2326/08/23
OtherInternational Conference on AI-generated Content, AIGC 2023 (25/08/2023-26/08/2023, Shanghai)

Keywords

  • Comic Illustrations
  • ControlNet
  • Diffusion Models
  • Multiple Component

ASJC Scopus subject areas

  • General Computer Science
  • General Mathematics

Fingerprint

Dive into the research topics of 'Divide and control: generation of multiple component comic illustrations with diffusion models based on regression'. Together they form a unique fingerprint.

Cite this