Abstract
The study fosters further developments in fuzzy pattern recognition by introducing concepts of Granular Computing and forming models of granular pattern recognition. Both unsupervised and supervised schemes of pattern recognition are discussed. A role of information granularity is outlined, especially in the context of characterization of classification results whose quantification in terms of information granules helps endow them with some tangible confidence mechanisms. We show how granularity of classification results relates with the distribution of patterns in the feature space. We also investigate a collaborative processing environment (when dealing with individual collections of patterns) and highlight a role of information granules in the formation of mechanisms of collaboration. The diversity of granular results and reconciliation of findings is quantified in terms of resulting information granules.