Properties of our visual system are deeply related to visual properties of our environments. Early visual environments in particular have a significant influence on the development of the visual system. To understand the nature of early visual input available to humans, researchers have captured natural environments from the perspective of young infants using head-mounted cameras [Aslin, 2009; Jayaraman et al, 2013; Sugden et al, 2013]. These environments typically characterize input available to infants in a specific context - middle class North American homes and their surroundings. However, infants' environments are arguably different across various socio-economic and cultural contexts. In this study, we captured the natural visual environments of twelve infants aged 0-14 months from a fisherman community of low socio-economic status in Chennai, India. We compared the images captured by these infants with images from age-matched infants from middle class homes in Bloomington, IN (USA). To ensure directly comparable images we used the same cap-mounted camera setup on both sets of participants. The resulting images reveal that visual environments of Indian infants are indeed very different from those of US American infants on several measures. Preliminary examination suggests that images from homes in India have lower luminance and higher contrast than those in the US. The range of hues is also narrower in the Indian images. These differences could be a function of geographic (quality of natural light), economic (availability of artificial lighting), and socio-cultural (aesthetic preferences) factors. While the factors that cause these differences are interesting in their own right, the implications of these differences are profound. Basic attributes of early input, such as luminance and contrast levels, are known to play a major role in the development of the visual system. If these attributes differ could that imply significant differences in the development of visual systems across populations?
Meeting abstract presented at VSS 2014