Partial matching in fuzzy image recognition

Abstract

The problem of how to build expert systems for recognizing vague or incomplete images is addressed. The focus is on the partial match and estimation of supports, per degree of similarity, of each possible option. Two methods were used: a voting model interpretation of fuzzy sets to classify the objects, and the iterative assignment algorithm to obtain the support. To test the learning capabilities of this approach, an example of upper case letters was implemented.<>