WAYS OF FEATURING IN OBJECT CATEGORIZATION

The origin of features from nonfeatural information is a problem that should concern all theories of object categorization and recognition, not just the flexible feature approach. In contrast to the idea that n ew features must originate from combinations of simpler fixed features , we argue that holistic features can be created from a direct imprint ing on the visual medium. Furthermore, featural descriptions can emerg e from processes that by themselves do not operate on feature detector s. Once acquired, features can be decomposed into component features i f required by other categorizations. We therefore argue that it is not necessary to separate holistic and componential approaches to represe ntations, because the latter is a development of the former. The requi rements for representational flexibility outstrip the performance of a ny existing computational models, but specific mechanisms of feature c reation are discussed and evaluated. Challenges for feature creation m echanisms are discussed together with the constraints (perceptual, sta tistical, functional, and task) they will need to satisfy.