Coordinate transformations in object recognition.

Abstract

A basic problem of visual perception is how human beings recognize objects after spatial transformations. Three central classes of findings have to be accounted for: (a) Recognition performance varies systematically with orientation, size, and position; (b) recognition latencies are sequentially additive, suggesting analogue transformation processes; and (c) orientation and size congruency effects indicate that recognition involves the adjustment of a reference frame. All 3 classes of findings can be explained by a transformational framework of recognition: Recognition is achieved by an analogue transformation of a perceptual coordinate system that aligns memory and input representations. Coordinate transformations can be implemented neurocomputationally by gain (amplitude) modulation and may be regarded as a general processing principle of the visual cortex.Keywords: alignment, coordinate transformations, gain modulation, object recognition, reference frames How can we recognize objects regardless of spatial transformations such as plane and depth rotation, size scaling, and position changes? This ability is often discussed under the label object constancy or shape constancy. Even young children recognize objects so immediately and effortlessly that it seems to be a rather ordinary and simple task. However, changes in the spatial relation between observer and object lead to large changes of the image that is projected onto the retina. Hence, to recognize objects regardless of orientation, size, and position is not a trivial problem. No computational system proposed so far can successfully recognize objects over wide ranges of object categories and contexts.Several different approaches have been proposed over the years (for reviews, see Palmeri & Gauthier, 2004;Ullman, 1996). A number of models rely on abstract object representations, which predict that recognition performance is typically invariant regarding spatial transformations (e.g., structural description models; see Hummel & Biederman, 1992;Marr & Nishihara, 1978). In contrast, image-based or view-based models propose that object representations are close to the format of the perceptual input and therefore depend systematically on image transformations (e.g.,

A stereo disadvantage for recognizing rotated familiar objects

“…On the face of it, this result seems counterintuitive, given that mono stimuli contain a subset of the information present in the stereo condition. Specifically, the key finding in our experiments is that the cue of outline shape seemed to assist performance in Experiments 1 and 2 only in the mono condition; recognition of 180º rotations in the stereo condition of these experiments seemed to be based on a normalization mechanism more closely aligned with the spatial orientation of the object (e.g., Graf, 2006;Tarr, 1995).…”

Rotating With Rotated Text: A Natural Behavior Approach to Investigating Cognitive Offloading

“…However, numerous studies have questioned the notion that the cost associated with identifying a rotated object is due to the need to mentally rotate an internal representation of that object (Cheung, Hayward, & Gauthier, 2009;Gauthier et al, 2002;Jolicoeur, Corballis, & Lawson, 1998;Wilson & Farah, 2006). An alternative account of the alignment process views it as a coordinate transformation (Graf, 2006). According to this account, alignment consists of the analog rotation of an internal reference frame to match the stimulus.…”

“…For example, presenting objects in a non-canonical orientation is often used in an attempt to understand constancies in visual object recognition (e.g., Tarr, 1995). There exist numerous demonstrations that performance, across a range of tasks, can be impaired when stimuli are not presented in their canonical orientation (e.g., Corballis, 1988;Graf, 2006;Jolicoeur, 1990a;Kung & Hamm, 2010). This appears to be true for a number of different classes of stimuli and tasks, for example, naming line drawings of natural objects (Jolicoeur, 1985), naming single letters (Jolicoeur & Landau, 1984;Jolicoeur, Snow, & Murray, 1987;Kolers & Perkins, 1969a, b), reading words (Koriat & Norman, 1984), and recognizing scenes (Diwadkar & McNamara, 1997).…”

“…Understanding how individuals recognize objects that are not presented in their canonical orientation (i.e., rotated objects) has long represented a fundamental problem in cognitive psychology (e.g., Corballis, 1988;Graf, 2006;Jolicoeur, 1990a;Tarr, 1995;Tarr & Pinker, 1989). For example, presenting objects in a non-canonical orientation is often used in an attempt to understand constancies in visual object recognition (e.g., Tarr, 1995).…”

Information Routing, Correspondence Finding, and Object Recognition in the Brain

“…Another point that begs discussion is the important role of priming or congruency effects in general object perception (for a review, see Graf 2006). For example, when we look at the moon rising over a distant skyline, it looks much larger than when we see it high in the sky.…”