From my personal experience, I do not think darker surfaces are regarded preferably as the object, and light surfaces as the background.
I have been pondering on this question for so long that I decided to post an answer. Perhaps it will draw attention from others who can give a more definitive answer.
This answer was inspired by this question on the Rubin Vase Illusion. It is an ambiguous figure/ground illusion which is thought to depend on the visual system analyzing scenes in terms of object (figure) and background (ground). You can see a vase when you regard the faces as background and vice versa. In the Fig. 1 below I have posted two versions of this illusion where the black and white portions are reversed. Personally, and hence anecdotally, I do not perceive a clear difference in the perception of the vase versus faces between both figures. Hence, based on this anecdotal evidence I tend to believe that the definition of figure and ground is independent on color (shades of gray in this case). I couldn't find references for the flip-flop rates or vase-versus-face perception time on the two stimuli, unfortunately. In all, what a great question!
Fig. 1. Contrast-inverted Rubin Illusions. sources: Winkler et al. (2006) (left panel) and Russel & Dowey (2014) (right panel)
- Russel & Dowey, Psychology: An Introduction (2014)
- Winkler et al. Eur J Neurosci (2006); 24(2): 625–34