Just as in mental rotation experiments gradual and monotonic change in viewpoint precipitates an equally gradual and monotonic change in the performance, the error rate and the response time for a test view of a 3D shape was found to grow monotonically with misorientation relative to a canonical view.
They showed that, when viewing pictures of objects of different physical sizes within a frame, smaller sizes within the frame were preferred for smaller objects in the real world (eg strawberries or a key), whereas larger sizes in the frame were preferred for larger real-world objects (eg a piano or chair). They called these effects `canonical size' in analogy with Palmer et al's (1981) `canonical perspective' effects, showing that people systematically prefer some perspective views of objects over others.
Our results suggest that canonical views are explicable by geometric factors such as robustness against rotation. Therefore, the results may also be applicable to common objects. The canonical views in our results were the perspectives in which the 3-D structure of the object could be clearly perceived. Such structural information may play some role even in the recognition of paper-clip objects.
With regard to our present results, the heavy weighting of familiarity in determining canonicality suggests that observers are relying more on their experience with particular views than on object geometry or functional considerations. Such a pattern is unlikely to arise if views are determined purely by encounters with previously-unseen geometric configurations (Biederman & Gerhardstein, 1993). In contrast, this is precisely the pattern predicted by theories of recognition in which views are determined by statistical learning methods over repeated exposures (Poggio & Edelman, 1990).
Most recently Khalil and McBeath reported the results of a study in which they explicitly asked their participants to rate their aesthetic judgments of different perspective views. These aesthetic preferences generally corresponded well with the results reported by Palmer, Rosch, and Chase. Ecological perspective biases associated with canonical perspective thus reflect another way in which people's aesthetic preferences reveal implicit knowledge about objects in the world: people like pictures of objects that make them most recognizable by showing their most informative parts and interrelations.
One explanation of canonical viewpoints that is consistent with both viewpoint-specific and viewpoint-irrelevant theories is that they represent a frequency effect. Tarr and Pinker (1989), for example, suggested that canonical viewpoints for novel objects form only after repeated viewings at a specific orientation. Karnath, Ferber, and Bülthoff (2000) suggested that the neurons that underlie the recognition of an object become attuned to the features present in the most recognizable views of that object, and that this tuning evolves with experience.
Are there similar preferences for particular views of scenes? We investigated this question using panoramic images, which show a 360-degree view of a location. Observers used an interactive viewer to explore the scene and select the best view. We found that agreement between observers on the “best” view of each scene was generally high.
It seems to be a universal trait that we think about, remember, imagine and recognize objects from this canonical perspective. Why care? Well, if you want to use icons at your web site or in your web or software application that people will recognize, then you might want to use this perspective. This is probably not so critical if you are using a well known logo, for example, the logo for itunes or Firefox, but becomes important if the icon is not as familiar, such as recognizing below that one of the logos is of a truck, or a photo printer.
Objects are recognized faster when presented from a canonical perspective ("three-quarters view"; Blanz, Tarr, & Bülthoff, 1999). Moreover, the more disoriented an object is with respect to the canonical perspective, the more time it takes to identify the object (Bülthoff, Edelman, & Tarr, 1995; Tarr & Bülthoff, 1998). For novel objects, new perspectives may become canonical, if subjects are presented repeatedly with these views.
When we identify an object, there is a preferred canonical perspective that contains the most salient characteristics for recognizing that particular object. This "prototypical" perspective implies the existence of an internal category in memory for mapping the object, or the "necessary precondition for definition of the object and its attributes" (Palmer, Rosch, & Chase, 1981).