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: Directional GAN: A Novel Conditioning Strategy for Generative Networks

Abstract: Image content is a predominant factor in marketing campaigns, websites and banners. Today, marketers and designers spend considerable time and money in generating such professional quality content. We take a step towards simplifying this process using Generative Adversarial Networks (GANs). We propose a simple and novel conditioning strategy which allows generation of images conditioned on given semantic attributes using a generator trained for an unconditional image generation task. Our approach is based on modifying latent vectors, using directional vectors of relevant semantic attributes in latent space. Our method is designed to work with both discrete (binary and multi-class) and continuous image attributes. We show the applicability of our proposed approach, named Directional GAN, on multiple public datasets, with an average accuracy of 86.4% across different attributes.
Comments: Accepted to AICC workshop at CVPR 2021
Subjects: Computer Vision and Pattern Recognition (cs.CV); Neural and Evolutionary Computing (cs.NE)
Cite as: arXiv:2105.05712 [cs.CV]
  (or arXiv:2105.05712v2 [cs.CV] for this version)

Submission history

From: Shradha Agrawal [view email]
[v1] Wed, 12 May 2021 15:02:41 GMT (7842kb,D)
[v2] Thu, 13 May 2021 22:04:31 GMT (7842kb,D)

Link back to: arXiv, form interface, contact.