Instead of consisting of a single limb, it looks more like it was originally a three-pronged object, one of whose limbs has been removed. However, a subtle change in the object’s shape-roughening the concavity on the right side-completely changes the interpretation of the causal history of the object (bottom row). The shape in the top row appears to be a gently curved object consisting of a single trunk. 1).įigure 2 further illustrates the deep connection between shape representation and causal inference. Although Leyton’s framework is potentially powerful in capturing the visual representation of causal origin, it cannot explain differences in causal attribution for objects that are geometrically very similar ( Fig. Consequently, the presence of an extremum indicates that a force has acted on a shape and its direction indicates the direction of that force. These axes are defined by and extend into the curvature extrema along the contour (e.g., into the tips of the cookie or croissant in Fig. Specifically, forces that shape an object are thought to operate along local symmetry axes within the shape contour. He provided a framework in which simple shape features are interpreted in terms of causal attribution. Leyton 35, 36 was one of the first to suggest that this inference of causal origin plays a role in the visual perception of shape (also see ref. Somehow, by combining perceptual shape computations with experiences from our past, we are able to infer the causal origins of the two different kinds of concavity. The cookie is bitten while the croissant is bent. 1: despite the geometrical similarity of the 2D shapes, we readily assign different physical causes to their concavities. Consider, for example, the cookie and croissant in Fig. Here, however, we are interested in how shapes are perceived and represented depending on higher-level inferences about the causal origin of objects and their features. For example, to work out if an object is symmetrical, it is necessary to compare how corresponding object features co-vary relative to a symmetry axis, while factoring out other potential match locations 19, 20.Īlmost all work to date has focused on how perceptual organization depends on the geometrical properties of shapes, for example, when explaining how shapes are decomposed into parts 21, 22, 23 or how missing pieces of occluded or fragmented shapes are interpolated 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34. In order to make higher-level inferences about objects (e.g., identifying symmetry axes, identifying front and back ends, predicting likely physical weight, or parsing the object into functional parts, like the cup and handle of a mug), the brain must somehow pool and organize information from distant locations across the object into more global quantities-a process known as perceptual organization 17, 18. It is clear that there is much more to shape perception than making local judgments about surface structure. To make matters worse, inferring the geometrical properties of local surface patches-such as their depth, orientation or curvature-is only the start of the shape-inference process. However, estimating and representing object shape from the retinal image is far from trivial 1, 3, 4, 12, 13, 14, 15, 16. Shape perception is crucial for many tasks including object recognition 1, 2, 3, 4, guiding reaching and handling actions 5, 6, 7, 8, and making other high-level inferences about object properties, such as whether an object is physically stable or likely to topple over 9, 10, 11. One of the most important functions of vision is to infer and represent the shape of objects. Objects are not only parsed according to what features they have, but also to how or why they have those features. The findings suggest that visual shape representations are more sophisticated than previously appreciated. This suppression of bitten regions was also found when observers were not asked about symmetry axes but about the perceived front and back of objects. However, when objects appeared ‘bitten’-as if parts had been removed by a distinct causal process-the responses deviated significantly from the geometrical axes, as if the bitten regions were suppressed from the computation of symmetry. When objects appeared ‘complete’-created entirely by a single generative process-responses closely approximated the object’s geometrical axes. Observers placed dots on objects to report their perceived symmetry axes. Here, however, we find that shape representations are also profoundly influenced by an object’s causal origins: the processes in its past that formed it. Most theories of shape perception focus exclusively on geometrical computations (e.g., curvatures, symmetries, axis structure). One of the main functions of vision is to represent object shape.
0 Comments
Leave a Reply. |