abstract: In this talk we will present a discussion about perceptual-based color correction in digital images in the framework of variational techniques. The fundamental difference between global and local enhancement is explored and linked to the variational formulation of a particular model of color enhancement: ACE. The variational point of view allows to put in evidence many properties of ACE which are hidden in its basic formulation. Finally, we also show that the gradient descent technique to minimize the ACE energy functional (and other related versions,) combined with a Taylor expansion of the related Euler-Lagrange equation, reduces the computational complexity of ACE from O(N2) to O(N log N). In practice this means that high quality perceptual-based color corrected versions of high resolution images may be obtained in a matter of seconds, instead of tens of minutes, with a regular PC.