From my experience:
Use 3 solid layers on the bottom, so the brightest part of the image will have that much under it to support it at a minimum, for strength.
Use 100% infill for the remaining layers.
Depending on the transparency of your filament, distribute the image on top of the bottom layers over about 30 layers. You want the highest number of layers you can live with to give the maximum number of different levels of brightness. Thinner layers will let you use more of them.
If you use too many layers the image gets dark and contrast goes down.