[Re] Exact Feature Distribution Matching for Arbitrary Style Transfer and Domain Generalization

Reproducibility Study

In this reproducibility study, we present our results and experience during replicating the paper, titled Exact Feature Distribution Matching for Arbitrary Style Transfer and Domain Generalization [1]. In real‐world scenarios, the feature distributions are mostly much more complicated than Gaussian, so only mean and standard deviation may not be fully representative to match them. This paper introduces a novel strategy to exactly match the histograms of image features via the Sort‐Matching algorithm in a computa‐ tionally feasible way. We were able to reproduce most of the results presented in the original paper both qualitatively and quantitatively.

Bibtex:

@inproceedings{erkol2023re,
  title={[Re] exact feature distribution matching for arbitrary style transfer and domain generalization},
  author={Erkol, Mert and K{\i}nl{\i}, Furkan and {\"O}zcan, Bar{\i}{\c{s}} and K{\i}ra{\c{c}}, Furkan},
  booktitle={ML Reproducibility Challenge 2022},
  year={2023}
}

Original Paper:

@inproceedings{zhang2021exact,
  title={Exact Feature Distribution Matching for Arbitrary Style Transfer and Domain Generalization},
  author={Zhang, Yabin and Li, Minghan and Li, Ruihuang and Jia, Kui and Zhang, Lei},
  booktitle={CVPR},
  year={2022}
}