I took a more conventional approach, using selective colors and a variety of other tricks, eventually bringing me to this unsatisfactory point.

The red areas are too extensive to yield to the healing brush, and dodging won't really solve the problem. The color problem will still be there even if we get the values in sync.

Whenever I have a problem like this one, I like to treat it like a simple color correction issue. I have one area that I don't like, and another area that I'd like to match. It's a perfect problem for a set of curves to solve; the only problem is creating a decent mask through which to engineer the shift. My solution is to work backwards.

First, I set a color sampler in the middle of the triangular patch of skin just to the left of the nose, as the tone I wanted to match. Then I set a second sampler somewhere in the middle of the red blob just to the left of the first area. The red sampler read R:209, G:135, B:108. The clear sampler read R:229, G:182, B:156.

Next, set a curve adjustment layer and fill the layer mask with black. Make sure that the gradient at the bottom of the window has black on the left. now set a midpoint in each of the channels, plugging in the values as they appear in the info palette. In the red curve, with the midpoint selected, you would enter 209 for Input, 229 for Output. In the green curve use 135 for input, 182 for output, and in the blue curve use 108 for input, 156 for output.

I took a more conventional approach, using selective colors and a variety of other tricks, eventually bringing me to this unsatisfactory point.

The red areas are too extensive to yield to the healing brush, and dodging won't really solve the problem. The color problem will still be there even if we get the values in sync.

Whenever I have a problem like this one, I like to treat it like a simple color correction issue. I have one area that I don't like, and another area that I'd like to match. It's a perfect problem for a set of curves to solve; the only problem is creating a decent mask through which to engineer the shift. My solution is to work backwards.

First, I set a color sampler in the middle of the triangular patch of skin just to the left of the nose, as the tone I wanted to match. Then I set a second sampler somewhere in the middle of the red blob just to the left of the first area. The red sampler read R:209, G:135, B:108. The clear sampler read R:229, G:182, B:156.

Next, set a curve adjustment layer and fill the layer mask with black. Make sure that the gradient at the bottom of the window has black on the left. now set a midpoint in each of the channels, plugging in the values as they appear in the info palette. In the red curve, with the midpoint selected, you would enter 209 for Input, 229 for Output. In the green curve use 135 for input, 182 for output, and in the blue curve use 108 for input, 156 for output.