Visual ability in newborns is something of a contested topic (as are most in early childhood development) in Developmental Psychology. There are several different areas of research within this category, each with its own 'major' issues.
While physiologically newborns possess all the necessary structures to see color, research suggests that they cannot perceive the whole palette available to adults. A short review of pertinent literature reveals claims ranging from newborns possessing a rudimentary of system used by western adults to them perceiving virtually no color at all (Bornstein, 1976; Werner & Wooten, 1979). What is agreed upon is that by the age of two months, infants' color perception abilities are nearly equal to those of adults (Bornstein, 1988).
It is agreed that newborns are myopic (agreement? egads!). Unfortunately, having narrowed the debate to simply degree still provides ample area for argument. Current estimates range from 20/300 to 20/800 for newborn visual acuity (Fantz, Ordy, & Udelf, 1962; Cornell & McDonnel, 1986). In reality, even the worst case would make little difference to an infant in daily life, where the only demand made upon the newborn is making eye contact while nursing. Without the ability to locomote significantly or even hold up their own head this is hardly a major deficit. We do find that by the time that infants are able to crawl at 7 to 8 months of age, their acuity approaches that of adults (Haith, 1990; Cornell & McDonnel, 1986).
Oddly, newborns continually scan their environment, in spite of their lack luster visual acuity (Bronson, 1991; Haith 1980). Marshall Haith went so far as to develop a method of recording a newborn's eye movements in a completely darkened environment. In doing so, he discovered that the visual scanning continued even in complete darkness. Without any external stimuli to elicit this behaviour, it had to be generated by other neural activity. These endogenous eye movements were also complemented by primitive forms of exogenous looking, where the eye was drawn to changes taking place in their environs.
Up until the early 1960's, it was thought that infants simply saw light without perceiving forms. This notion was destroyed by Robert Frantz when he showed that newborns can distinguish among different visual forms with which they are presented (1961, 1963). Further research suggested that early on infants were most drawn to figures that exhibit high contrast and high complexity. It also suggested that when young, infants tend to focus only on portions of the stimuli scanning the objects presented in a more systematic manner as they age (Bronson, 1994).
While Fantz's early studies found a preference for faces in newborns, they failed to control for complexity of presented stimuli. Once this variable has been reined in, we see no preference for faces until about 12 weeks of age. It is also important to note the importance of newborn interest in movement, which is inevitably present in the human face. Finally, while there is some evidence to suggest that newborns only 2 days old can recognize the face they have seen most often, it is seeming ever more likely that this is little more than an ability on the part of the newborn to detect differences in hairline contours (Bushnell, Sai, & Mullin, 1989; Pascalis et al., 1995).
For a full bibliography, message me. It took up too much space here.