Robust Colour Transfer using the L2 Divergence



Summary

This work proposes to perform colour transfer by minimising the L2 distance between two colour distributions. We first propose to model each colour distribution as a compact Gaussian mixture which is designed for the specific purpose of performing colour transfer between images both with similar and different scene content. A non rigid transformation is estimated by minimising the L2 distance between these two distributions, and the estimated transformation is used to transfer colour statistics from one image to another. Experimental results show our technqiue outperforms other state of the art methods.

In [4], we present an initial exploration of this approach applied to images with different content, and extend the method to video content in [3]. In [1,5] we show that our approach can also be successfully applied to images of the same scene, taking into account pixel correspondences between images. We show that several colour transformations can be easily combined to create interesting visual effects in images. In [1] we also give a more in depth comparison between our technique and optimal transport recolouring methods. In [2], we show that this framework can also be extended to allow for user interaction in the recolouring process.



Publications

  1. L2 Divergence for robust colour transfer , M. Grogan, R. Dahyot. Computer Vision and Image Understanding. (CVIU), 2019
    PDF

  2. User Interaction for Image Recolouring using L2 , M. Grogan, R. Dahyot and A.Smolic. ACM SIGGRAPH European Conference on Visual Media Production (CVMP) , London, December 2017. Awarded Best Paper!
    PDF Video Webpage

  3. L2 Registration for Colour Transfer in Videos , M. Grogan, R. Dahyot. ACM SIGGRAPH European Conference on Visual Media Production (CVMP) , London, November 2015. Awarded Best Poster!
    PDF Video

  4. L2 Registration for Colour Transfer , M. Grogan, M. Prasad, R. Dahyot, European Signal Processing Conference (Eusipco) , Sept 2015.
    PDF

  5. Robust Registration of Gaussian Mixtures for Colour Transfer, M. Grogan and R. Dahyot, Arxiv, May 2017
    PDF


Online Demos

Try our online example based recolouring DEMO now! For more user interaction, try these DEMOS too!



Code

Our code is available on Github.





Some Recolouring Results


Original Image Palette Image Result
Home