Development of object recognition

Object recognition is the process by which humans organize the visual world into meaningful perceptual units. To understand this ability in humans, it is important to examine its origins in infancy and the processes by which it reaches maturity. In this review, we examine the development of object recognition by synthesizing research from developmental psychology, cognitive neuroscience, and computational modelling. We describe how, within the first year, infants demonstrate early traces of adult visual competencies, ranging from invariant object recognition to few-shot category learning. The rapid development of these competencies is supported by infant-specific biological and experiential constraints, such as low-visual acuity and innate biases for properties like symmetry. Furthermore, infants’ experience with objects is ‘self-curated’, such that they select object viewpoints that best support learning. Indeed, incorporating infant-like constraints into computational models improves their performance on many recognition tasks. The neural mechanisms that support these abilities in infancy, however, may differ from those in adulthood: whereas the ventral visual pathway is most crucial for object recognition in adults, object recognition in infants may be primarily supported by lower-level visual properties and, potentially, dorsal visual pathway representations. Together, these studies highlight the importance of children’s specific developmental niche in shaping early object recognition abilities and their neural underpinnings.