Skip to content

Topic 12: Color transfer

  1. Read and discuss the paper by Pitié et al..
  2. Understand iterative distribution transfer (IDT) summarized in “Algorithm 1”.Explain the gist of the method in the report.
  3. Test your implementation of IDT on two self-generated distributions and plot the intermediate results similar to “Fig. A.1.” in the paper.
  4. Describe and show the grain noise artefact in color transfer mapping.
  5. For implementation details, read carefully section “4.1. Numerical solution”.
  6. Capture your own image pairs and try to perform color transfer on them.

Typo

In the paper by Pitié et al. there is an error from equation 16 to 17. They have interchanged \phi and \psi. That is also the case in the following equations for the coefficients a_1 to a_6. Please take this into account.

Equation 17 correctly reads

\psi \cdot J - div(\phi \cdot \nabla J) = \psi \cdot t(I) - div (\phi \cdot \nabla I)

In addition the coefficient a_6 has a typo, missing t()in the last term and should read

a_6 = \dots +\psi(x,y)\cdot t(I(x,y))

References

François Pitié, Anil C. Kokaram, and Rozenn Dahyot. 2007. Automated colour grading using colour distribution transfer. Comput. Vis. Image Underst. 107, 1–2 (July, 2007), 123–137.

Last update: 2024-05-23