← New search

Other meanings of Recognition-by-components theory

Cognitive Psychology

Recognition-by-components theory

Recognition-by-components theory (RBC) is a cognitive psychology theory proposed by Irving Biederman in 1987 to explain how humans recognize objects from their component parts. It posits that objects are perceived as arrangements of simple three-dimensional shapes called geons (geometric ions), which serve as the basic building blocks of object perception. The theory is influential in vision science and has applications in computer vision and object recognition.

1987
Year proposed
Year the theory was introduced by Irving Biederman
36
Number of geons
Number of basic geons proposed in the theory
~30,000
Recognizable objects
Estimated number of objects that can be recognized using geon combinations
1

Core principles

Recognition-by-components theory holds that object recognition proceeds by decomposing an object into its constituent geons, which are viewpoint-invariant primitives such as cylinders, cones, and blocks. Biederman proposed that there are 36 such geons, defined by non-accidental properties (NAPs) like edge parallelism, collinearity, and curvature, which remain stable across different viewing angles. The theory suggests that the visual system extracts these geons from the retinal image and matches them against stored representations, enabling rapid and robust recognition even when objects are partially occluded or degraded.

2

Evidence and support

Empirical studies have supported RBC by showing that recognition performance is better when geon information is preserved than when it is degraded. For instance, Biederman and colleagues found that removing contour information that defines geons (e.g., by deleting vertices) impairs recognition more than removing non-geon information. Neuroimaging studies have also identified brain regions, such as the lateral occipital complex, that respond selectively to geon-like features. However, the theory has been critiqued for underestimating the role of context, familiarity, and viewpoint-dependent processes in object recognition.

3

Applications and influence

RBC has influenced computer vision, particularly in the development of object recognition algorithms that use geon-like features for 3D object modeling. It has also informed theories of visual perception in artificial intelligence, where geons are used as primitives for shape representation. In psychology, RBC has been applied to understand visual agnosia, a condition in which patients lose the ability to recognize objects despite intact vision, suggesting that damage to geon extraction mechanisms may underlie such deficits.

4

Lesser-known aspects

Beyond its mainstream applications, RBC has been extended to explain recognition of faces and scenes, though with modifications. Biederman also proposed that geons can be combined in a combinatorial manner, yielding a vast number of possible objects, similar to how words are formed from letters. The theory has been tested in cross-cultural studies, showing that geon-based recognition is universal across different visual environments. Additionally, RBC has been used in the design of tactile graphics for visually impaired individuals, where geon-like shapes are used to convey object structure through touch.

Glossary

Geon
A simple three-dimensional shape, such as a cylinder or cone, that serves as a basic building block for object recognition.
Non-accidental properties (NAPs)
Visual features that are invariant to viewpoint, such as parallelism and collinearity, used to define geons.
Viewpoint invariance
The property that an object's recognition is unaffected by changes in viewing angle.

This article is based on peer-reviewed literature and authoritative sources.