APPLICATION OF FUZZY LANGUAGES TO PATTERN RECOGNITION

Abstract

The concept of a fuzzy language is applied to pattern recognition using geometric figures, chromosomes and leukocytes as illustrative examples. For chromosomes, an algorithm for classifying a chromosome image as an “approximate median chromosome,” “approximate sub‐median chromosome” or “approximate acrocentric chromosome” is presented. For leukocytes, an equal‐perimeter circular shape measure and an equal‐area circular shape measure are proposed, and various properties, results, and the relationship between these two measures are presented. Quantitative measures of other visual concepts such as elongated, spiculed, indented, slightly indented and deeply indented arc also presented and illustrated by examples. The results obtained in this paper may have useful applications in pattern recognition, cybernetics and fuzzy systems.