Abstract
Everyday objects are more readily recognized when seen from certain representative, or canonical, viewpoints than from other, random, viewpoints. We investigated the canonical views phenomenon for novel 3D objects. In particular, we looked for the effects of object complexity and familiarity on the variation of response times and error rates over different views of the object. Our main findings indicate that the response times for different views become more uniform with practice, even though the subjects in our experiments received no feedback as to the correctness of their responses. In addition, the orderly dependency of the response time on the distance to a good view, characteristic of the canonical views phenomenon, disappears with practice. One possible interpretation of our results is in terms of a tradeoff between memory needed for storing specific-view representations of objects and time spent in recognizing the objects.
