Infants’ identification of three-dimensional form from transformations of linear perspective

Abstract

The ability of infants to perceive three-dimensional structure from transformations of linear perspective was investigated in two studies. Infants were habituated to the pattern oflinear perspective transformations corresponding to a particular three-dimensional object, and their relative preference for that object as compared with a different three-dimensional object was assessed both before and after habituation. The habituation displays showed the distorting shadow cast by a rotating object and therefore provided only transformations of linear perspective as information specifying three-dimensional form. The pre-and posttest displays involved the actual threedimensional objects and provided binocular, shading, and texture information specifying threedimensional form, but did not provide informative transformations oflinear perspective. In Study 1, 6-month-olds showed changes in preference from pre-to posttest that were related to the identity of the object whose shadow they had seen during habituation; 4-month-olds, however, did not show preference changes related to the habituation object. In Study 2, rhythm information that may have served as a basis for responding in Study 1 was eliminated from the test displays. Sixmonth-olds again showed changes in preference that were dependent on their habituation experience. It is concluded that, by 6 months of age, infants are able to perceive object structure from the isolated cue of transformations of linear perspective. The findings are discussed with reference to infants' three-dimensional form perception based on other cues and also with reference to the emergence of certain spatially related moter activities.The ability to perceive the three-dimensional structure of objects and their spatial layout in the environment is essential for effective motor behavior; consequently, depth perception has been a matter of considerable interest for psychologists. The central problem with respect to this matter is that the perception of three-dimensional objects in a three-dimensional space must be derived from retinal images that are represented two-dimensionally. Psychologists have identified a variety of "cues" to depth that apparently mediate this conversion process and, to determine the origins and development of depth perception, the responses of young infants to a number of these cues have been investigated. The evidence, recently reviewed by Banks and Salapatek (1983), indicates that infants are, indeed, sensitive to many of the identified cues to depth by as early as 2-3 months of age. Furthermore, observations of infants' reaching behavior indicate that actual distance relationships (nearness/farness) specified by proximal cues are appreciated by 5 months of age in the case of disparity cues (Gordon & Yonas, 1976) and

Development of three-dimensional form perception.

“…Recently, Shaw, Roder, and Bushnell (1986) argued in regard to an earlier presentation of this study (Kellman & Short, 1985) that infants might have discriminated the two objects in this study without perceiving 3-D form, by detecting for each object the angles of intersection of its component triangles at the moment in its rotation when its projection consists of only two lines. Their argument misunderstands the design of this study, in which this particular view (or any 2-D view) of an object occurred either in habituation or in the test trials but never in both.…”

“…Shaw and Bushnell (1983) suggested that changes in brightness and texture are necessary for early 3-D form perception. Using a shadow-casting technique, they displayed perspective tranformations of a rotating object in the absence of brightness and texture gradients.…”

Velocity thresholds in human infants: Implications for the perception of motion.

“…Moreover, movement leads to systematic motor responses such as optokinetic nystagmus (Hainline, Lemerise, Abramov, & Turkel, 1984; Kremenitzer, Vaughan, Kurtzberg, & Dowling, 1979; McGinnis, 1930; Tauber & Koffler, 1966), pursuit tracking (Aslin, 1981; Dayton, Jones, Steele, & Rose, 1964; Kremenitzer et al, 1979; Roucoux, Culee, & Roucoux, 1983), reaching (Hofsten, 1980, 1983; Hofsten & Lindhagen, 1979), and postural control (Bertenthal & Bai, 1989; Butterworth & Hicks, 1977; Lee & Aronson, 1974). Furthermore, stimulus movement imparts important information about the properties of objects, such as relative size and distance (Granrud, 1986), shape (Kellman, 1984; Kellman, Hofsten, & Soares, 1987; Kellman & Short, 1987; Shaw, Roder, & Bushnell, 1986; Yonas, Arterberry, & Granrud, 1987), the connectedness of object components (Hofsten & Spelke, 1985; Kellman & Spelke, 1983; Kellman, Spelke, & Short, 1986; Spelke, Hofsten, & Kestenbaum, 1989), and the geometric structure of point–light displays (Bertenthal, Proffitt, Kramer, & Spetner, 1987; Fox & McDaniels, 1982). Finally, stimulus movement is ubiquitous in that spontaneous eye, head, and body movements create nearly continuous motions of the image of the visual environment across the retina.…”

Infants’ sensitivity to kinetic information for three-dimensional object shape

“…In addition, the motion of objects in the environment provides kinetic optical information for their shape and motion in depth. The results of recent studies suggest that infants in the first year of life are able to perceive some aspects of three-dimensional object shape specified by kinetic information (Kellman, 1984;Kellman & Short, 1985;Owsley, 1983;Shaw, Roder, & Bushnell, 1986;Yonas, Arterberry, & Granrod, 1987). The goal of the present study was to extend this work by demonstrating that 4-month-olds can perceive the difference between two rectilinear forms specified by kinetic information in the absence of figural features such as contours and intersections.…”