---
tags:
- color-transfer
pretty_name: Color Transfer Publication Dataset
---
# Color Transfer Publication Dataset
The evaluation of color transfer algorithms presents challenges due to the diverse image sets used across different publications to demonstrate their efficacy. The inherently subjective nature of color transfer evaluation compounds this issue, as there isn't a universally accepted objective metric that flawlessly gauges result quality. To standardize evaluations and enable more accurate comparisons between algorithms, it's crucial to assess each algorithm using a consistent set of images. This dataset addresses this need by providing source, reference, and output images from a range of publications.

## Details
- Includes Source, Reference, and Output images curated from past color and style transfer studies.
- Facilitates streamlined comparison and validation of outcomes from various publications.
- Features images from 13 distinct studies (refer to the README within the dataset for specific publications).
- Images have been post-processed to eliminate any labels.
- Only images with the appropriate license are available.

<p align="center">
  <img src="https://cdn-uploads.huggingface.co/production/uploads/6331b00ab7b8e9d6e89a5dc0/nQNEM1PgQuLsgDra2OpCi.webp" width="512">
  <em>Figure 1: Example images from the dataset.</em>
</p>