Recognition-by-components (RBC) theory (Biederman, 1987) has greatly influenced our understanding object perception and, more specifically, recognition. Recognition has been defined by Goldstein (2010) as how humans can categorise objects to give them meaning. RBC suggests that our recognition of objects is dependent on dividing the stimulus into a number of “geons” (geometric ions), which are defined by the contour (edge) information. These geons are various 3-dimensional objects such as cylinders and blocks. This theory has numerous strengths and weaknesses which have been pointed out by supporters …show more content…
The collection of geons, combined with the spatial relationships and orientation of geons represent objects in the visual field (Harris, 2014). This theory assumes that the representation of the stimulus can be divided into separate areas, particularly when there is discontinuity, as suggested by Marr and Nishihara (1978). This theory was a development of the theory that the brain views things in three stages (Marr, 1982), starting as just a “primal sketch” with some 2-dimensional information, then adding binocular cues and other depth information before finally becoming a 3-dimensional model. Biederman (1987) uses the analogy of recognising language due to the phonemes of which it is composed to explain his theory, with the phonemes representing the geons and the language representing the object. This analogy is also used to explain how various geons can be rearranged to make a different object. Like phonemes being rearranged to make a new word, different objects can be composed of the same geons. The theory also suggests that the brain infers that if an object appears to be a straight line, for example, then it must be so. The visual system does not take into account the chance that a curved surface may appear to be a straight line due to the orientation and alignment of the eye and the edge (called an “accidental view”) as it is, according to Biederman, highly unlikely. …show more content…
This strengthens the theory because it shows how one geon description can account for most viewpoints (excluding accidental viewpoints) and explains why the visual system can recognise objects from almost any angle. According to Biederman (2000), this is due to the invariant contour properties found in geons (also knows as non-accidental properties, or “NAPs”). These NAPs include the curvature of the main axis, the expansion of the cross-section, the positive and negative curvature of the sides, convergence to the vertex and the cross-sectional change compared with aspect ratio (Amir, Biederman & Hayworth, 2012). This means that one geon can represent an object from all possible viewpoints. Another advantage is that so few geons are needed to represent a large number of objects. This “economy of representation” means memory storage is not overloaded by the need to recreate 3-dimensional objects (Hummel & Biederman, 1992). Because of the seemingly infinite number of combinations that can be made from a small number of geons, RBC creates an incredibly efficient way of understanding object recognition which removes the issue of memory storage. With regard to the studies conducted by Biederman (1987) and Biederman and Gerhardstein (1993), it has been found that participants tend to show an almost-perfect ability to recognise objects as long as geons are visible, regardless of visual noise. This is a