Perceptual Categorization of Cat and Dog Silhouettes by 3- to 4-Month-Old Infants

Abstract

Given evidence that silhouette information can be used by adults to form categorical representations at the basic level, four experiments utilizing the familiarization-novelty preference procedure were performed to examine whether 3-and 4-month-old infants could form categorical representations for cats versus dogs from the perceptual information available in silhouettes (e.g., global shape and external outline). Experiments 1 and 2 showed that infants could form individuated categorical representations for cat and dog silhouettes, whereas Experiments 3 and 4 revealed that infants could use silhouette information from the head, but not the body, to categorically separate the two species. These results indicate that general shape or external contour information that is centered about the head is sufficient for young infants to form individuated categorical representations for cats and dogs. The data thus provide information regarding the nature of the perceptual information that can be used by infants to form category representations for individual animal species and are discussed in terms of domain-general versus domain-specific processing accounts.

Transfer to Intermediate Forms Following Concept Discrimination by Pigeons: Chimeras and Morphs

“…Following longer viewing, dishabituation was observed with novel body cues as well as novel head cues. Preferential responding to head cues was also found with silhouettes of the stimuli by Quinn, Eimas, and Tarr (2001). There are two possible interpretations of this result: it could derive from objective properties of the parts of the stimuli-there might simply be more information present in the head partor it could derive from a preference for looking at faces, which is evident in human infants from a very young age (e.g., Valenza, Simion, Cassia, & Umilta, 1996).…”

Early development of perceptual expertise: Within-basic-level categorization experience facilitates the formation of subordinate-level category representations in 6- to 7-month-old infants

“…In addition, the category representation of humans by infants appears to be based on the overall structure of the stimuli-for example, a head region adjoining an elongated body axis with skeletal appendages (Quinn, 2004). The finding that the human representation is based on global information contrasts with results showing that representations for nonhuman animal species are based on part or attribute (i.e., featural) information-heads, in the case of cats versus dogs (Quinn, Eimas, & Tarr, 2001). These data on how infants represent humans correspond with the finding that expert object recognition by adults is characterized by holistic-configural processing (Gauthier & Tarr, 2002; but see Robbins & McKone, 2007).…”

“…The pictures were selected to be of approximately the same size. As a result, the infants were required to use cues such as shape, parts, or the pattern of correlation across the parts-cues that have been shown to be diagnostic of category membership for both infants and adults (Quinn et al, 2001;Rosch et al, 1976;Younger, 1985).…”

Learning an Alphabet of Shape and Appearance for Multi-Class Object Detection

“…Deselaers et al 2005;Epstein and Ullman 2005). But shape is also an important cue for object categorization, for instance humans do use shape by means of the objects silhouette to distinguish between categories even in early vision (Quinn et al 2001). Using shape instead of appearance is not novel but is less explored for the task of categorization.…”