Patternmatching Theory


PATTERN/OBJECT RECOGNITION THEORY. The perception of shape/form, including figurai pattern and detail, is generally achieved by organisms through analysis of stimulus features from the sensory input [cf., the Hoffding step/phenomenon - named after the Danish philosopher and psychologist Harald Hoffding (1843-1931), is the Gestalt-ist's term for the mental step through which the perception of an image makes contact with a memory trace]. Contour and edge perceptions are hypothesized to take place at the retinal level, and some vision experts propose that contour and edges are the basis of complex form perception. An information-processing theory or analysis of vision requires an initial stage of figural synthesis, which is the way that stimulus information is transferred from the icon and synthesized into a form. In order for pattern/shape recognition to occur, the synthesized information is transferred, subsequently, to memory to produce a unique response (cf., pattern-matching theory - assumes that individuals can recognize differences and similarities between certain patterns actually presented to them and contain patterns deposited in their "memory store"). One of the major problems for pattern recognition and perception theorists is to understand how the organism consistently recognizes forms/shapes when they are presented in different sizes and retinal locations, are degraded by poor or "noisy" environmental conditions, and are partially outlined in cartoon or picture-like formats. The pandemonium model/theory was an early and influential computer model of pattern perception. In its simplest form, it was based on a number of perceiving elements called "demons" that were tuned to detect specific features (e.g., a straight line, a half circle, etc.). Each low-level demon that was activated "shouted out," and the higher level demons decided what stimulus was presented by sifting through the "wild uproar" or pandemonium (named after the capital of Hell in John Milton's book Paradise Lost). One general approach in pattern recognition theory is the feature extraction/ feature detection theory, which involves template-matching theory processes - and which assumes that various internal representations (i.e., "templates") of objects are stored in memory, and new stimuli are processed by comparing them with the templates until a match is found. However, as a theory of human pattern recognition, it is too simple and it cannot, for example, account for the ability to recognize that a, A, A, and a are all examples of the same letter. Most research on form/ shape recognition and perception includes basic visual functions concerning luminance distribution that produce lines or Mach bands, discriminable differences in forms, figural aftereffects, visual illusion changes due to unspecified cues, and estimation of the vertical orientation. F. Attneave and M. Arnoult studied the psychophysics of form and demonstrated that judgments of attributes of abstract forms could be related to stimulus domain features such as complexity and area. J. J. Gibson describes three-dimensional perception and suggests that object perception can be based only on form perception; he argues that features are important where what counts is not the form per se but the dimensions of variation of form. L. Zusne makes the terminological distinction that form is the more general term, and shape is more specific, even though the terms form and shape frequently are used interchangeably. B. Julesz hypothesizes that the "primitives" for object perception are units called textons that operate during the initial ("preattentive") stage of vis-ion. According to Julesz, texture formation in object perception is automatic and happens almost instantaneously. In her feature integration theory (FIT), A. Treisman suggests that the "primitives" of object perception operate in five stages. Treisman's FIT focuses on how different attributes such as shape, color, tex ture, and size are integrated into a simple object. However, Treisman does not spell out exactly how the process of feature combination works in her FIT. Another approach, I. Biederman's recognition by components (RBC) theory, is based on "primitives," also, but rather than being elementary properties such as color and shape, the "primitives" in RBC theory are volumetric primitives (such as spheres, cubes, and cylinders) called geons (for geometric ion). The basic idea behind RBC theory is that objects are recognized by perceiving their geons. According to the principle of componential recovery, one can easily recognize an object if its geons can be identified. The basic message of Biederman's theory is that if enough information is available to enable one to identify an object's basic geons, the perceiver will be able to identify the object. The theories by Julesz concerning textons, Treisman concerning FIT, and Bied-erman regarding RBC all have in common the notion that the perception of objects involves a number of stages, beginning with "primitives" and ending with the combination of primitives into the complete perception of an object. See also ARTIFICIAL INTELLIGENCE; CONCEPT LEARNING/CONCEPT FORMATION, THEORIES OF; GESTALT THEORY/LAWS; INFORMATION/INFORMATION-PROCESSING THEORY; MACH BANDS; PERCEPTION (I. GENERAL), THEORIES OF; PERCEPTION (II. COMPARATIVE APPRAISAL), THEORIES OF. REFERENCES

Hoffding, H. (1889). [Mediating mental step].

Vierteljahrsschrift Wissen schaftlicher Philosophie, 13, 420458.

Attneave, F., & Arnolult, M. (1956). The quantitative study of shape and pattern perception. Psychological Bulletin, 53, 452-471. Selfridge, O. (1959). Pandemonium: A paradigm for learning. In D. Blake & A. Uttley (Eds.), Proceedings of the symposium on the mechanization of thought processes. London: HM Stationery Office. Uhr, L. (Ed.) (1966). Pattern recognition: Theory, experiment, computer simulations, and dynamic models of form perception and discovery. New York: Wiley. Minsky, M. L., & Papert, S. (1969). Percep-trons. Oxford, UK/Cambridge, MA: M.I.T. Press. Zusne, L. (1970). Visual perception of form.

New York: Academic Press. Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton Mifflin. Julesz, B. (1981). Textons, the elements of texture perception, and their interactions. Nature, 290, 91-97. Biederman, I. (1987). Recognition-by-components: A theory of human image understanding. Psychological Review, 94, 115-147. Treisman, A. (1993). The perception of features and objects. In A. Baddeley & L. Weiskrantz (Eds.), Attention: Selection, awareness, and control. Oxford, UK: Clarendon.

Conquering Fear In The 21th Century

Conquering Fear In The 21th Century

The Ultimate Guide To Overcoming Fear And Getting Breakthroughs. Fear is without doubt among the strongest and most influential emotional responses we have, and it may act as both a protective and destructive force depending upon the situation.

Get My Free Ebook

Post a comment