Classification in noisy environments using a distance measure between structural symbolic descriptions

• Published in: IEEE transactions on pattern analysis and machine intelligence Vol. 14; no. 3; pp. 390 - 402
• Main Authors: Esposito, F., Malerba, D., Semeraro, G.
• Format: Journal Article
• Language: English
• Published: Los Alamitos, CA IEEE 01-03-1992; IEEE Computer Society
• Subjects: Applied sciences; Artificial intelligence; Computer science; control theory; systems; Distortion measurement; Euclidean distance; Exact sciences and technology; Instruments; Learning systems; Noise measurement; Pattern classification; Pattern matching; Pattern recognition; Pattern recognition. Digital image processing. Computational geometry; Phase noise; Working environment noise; Learning; Formal language; Computer vision; Image processing; Noise; Classification; Distance measurement; Structural model; Computer system; Predicate calculus; Pattern matching
• ISSN: 0162-8828; 1939-3539
• DOI: 10.1109/34.120333

A definition of distance measure between structural descriptions that is based on a probabilistic interpretation of the matching predicate is proposed. It aims at coping with the problem of classification when noise causes both local and structural deformations. The distance measure is defined according to a top-down evaluation scheme: distance between disjunctions of conjuncts, conjunctions, and literals. At the lowest level, the similarity between a feature value in the pattern model (G) and the corresponding value in the observation (Ex) is defined as the probability of observing a greater distortion. The classification problem is approached by means of a multilayered framework in which the cases of single perfect match, no perfect match, and multiple perfect match are treated differently. A plausible solution for the problem of completing the attribute and structure spaces, based on the probabilistic approach, is also given. A comparison with other related works and an application in the domain of layout-based document recognition are presented.< >