Wired Perspectives: Multi-View Wire Art Embraces Generative AI
Zhiyu Qu1,  Lan Yang2,  Honggang Zhang2,  Tao Xiang1,  Kaiyue Pang1,  Yi-Zhe Song1
1SketchX, CVSSP, University of Surrey
2School of Artificial Intelligence, Beijing University of Posts and Telecommunications

Accepted by CVPR 2024
[Paper]    [Code]    [bilibili]   




Multi-view art done generatively by DreamWire. We present DreamWire as the first system that takes user prompt for each view as input -- either via the expressive vehicle of text or image -- and produces 3D line sculptures showing distinct interpretations when viewed at different angles, i.e., multi-view wire art (MVWA). Each MVWA instance is created by optimising a set of randomly initialised Bézier curves towards the visual matching objective of a universal generative prior (diffusion models or ControlNet).
A great thought begins by seeing something differently, with a shift of the mind’s eye.
-- Albert Einstein

BibTeX
@inproceedings{qu2023dreamwire,
  title={Wired Perspectives: Multi-View Wire Art Embraces Generative AI},
  author={Qu, Zhiyu and Yang, Lan and Zhang, Honggang and Xiang, Tao and Pang, Kaiyue and Song, Yi-Zhe},
  booktitle={CVPR},
  year={2024}
}