• Computer Science, Mathematics
  • Published in ArXiv 2019

Approximating Human Judgment of Generated Image Quality

@article{Kolchinski2019ApproximatingHJ,
  title={Approximating Human Judgment of Generated Image Quality},
  author={Y. Alex Kolchinski and Sharon Zhou and Shengjia Zhao and Mitchell L. Gordon and Stefano Ermon},
  journal={ArXiv},
  year={2019},
  volume={abs/1912.12121}
}
Generative models have made immense progress in recent years, particularly in their ability to generate high quality images. However, that quality has been difficult to evaluate rigorously, with evaluation dominated by heuristic approaches that do not correlate well with human judgment, such as the Inception Score and Fréchet Inception Distance. Real human labels have also been used in evaluation, but are inefficient and expensive to collect for each image. Here, we present a novel method to… CONTINUE READING

Tables and Topics from this paper.

References

Publications referenced by this paper.
SHOWING 1-10 OF 26 REFERENCES

Improved techniques for training gans

  • Salimans, Tim, +9 authors Xi
  • In Advances in neural information processing systems,
  • 2016
VIEW 6 EXCERPTS
HIGHLY INFLUENTIAL

GANSynth: Adversarial Neural Audio Synthesis

Pros and cons of gan evaluation measures

  • Borji, Ali
  • Computer Vision and Image Understanding,
  • 2019

A note on the inception score

  • Barratt, Shane, Sharma, Rishi
  • arXiv preprint arXiv:1801.01973,
  • 2018

Generating Adversarial Examples with Adversarial Networks

VIEW 1 EXCERPT