DepthwiseGANs: fast training Generative Adversarial Networks for realistic image synthesis

Mkhuseli Ngxande, Jules-Raymond Tapamo, Michael Burke

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

1 Citation (Scopus)

Abstract

Recent work has shown significant progress in the direction of synthetic data generation using Generative Adversarial Networks (GANs). GANs have been applied in many fields of computer vision including text-to-image conversion, domain transfer, super-resolution, and image-to-video applications. In computer vision, traditional GANs are based on deep convolutional neural networks. However, deep convolutional neural networks can require extensive computational resources because they are based on multiple operations performed by convolutional layers, which can consist of millions of trainable parameters. Training a GAN model can be difficult and it takes a significant amount of time to reach an equilibrium point In this paper, we investigate the use of depthwise separable convolutions to reduce training time while maintaining data generation performance. Our results show that a DepthwiseGAN architecture can generate realistic images in shorter training periods when compared to a StarGan architecture, but that model capacity still plays a significant role in generative modelling. In addition, we show that depthwise separable convolutions perform best when only applied to the generator. For quality evaluation of generated images, we use the Fréchet Inception Distance (FID), which compares the similarity between the generated image distribution and that of the training dataset.

Original languageEnglish
Title of host publication2019 Southern African Universities Power Engineering Conference/Robotics and Mechatronics/Pattern Recognition Association of South Africa (SAUPEC/RobMech/PRASA) - Proceedings
EditorsElisha Didam Markus
Place of PublicationPiscataway NJ USA
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages111-116
Number of pages6
ISBN (Electronic)9781728103693, 9781728103686
ISBN (Print)9781728103709
DOIs
Publication statusPublished - 2019
Externally publishedYes
EventSouthern African Universities Power Engineering Conference/Robotics and Mechatronics/Pattern Recognition Association of South Africa 2019 - Bloemfontein, South Africa
Duration: 28 Jan 201930 Jan 2019
https://ieeexplore-ieee-org.ezproxy.lib.monash.edu.au/xpl/conhome/8698527/proceeding (Proceedings)
https://www.aconf.org/conf_168897.html (website)

Conference

ConferenceSouthern African Universities Power Engineering Conference/Robotics and Mechatronics/Pattern Recognition Association of South Africa 2019
Abbreviated titleSAUPEC/RobMech/PRASA 2019
Country/TerritorySouth Africa
CityBloemfontein
Period28/01/1930/01/19
Internet address

Keywords

  • Depthwise Separable Convolution
  • FID
  • GANs
  • Synthetic Data

Cite this