A Similarity Measuring Method Between Image Regions Based on Granular Computing
Jinling Shi, Genyuan Du · 2010
To effectively measure the similarity between image regions, this paper proposes a new measuring method based on granular computing theory. First, image feature information table is transformed into the type of order matrix by defining the concept of image feature information table, order matrix, feature granular and granular base. Secondly, further study and analysis of order matrix and feature granular with granular computing model are required. Moreover, a similarity measuring method is defined on condition that it doesn't change the image information based on the concepts of feature values and feature weights. Lastly, a detailed example is given to prove the validity of the similarity measuring method.