Affiliations: Department of Applied Computer Science, University of Winnipeg, MB, Canada
Correspondence:
[*]
Corresponding author: Sheela Ramanna, Department of Applied Computer Science, University of Winnipeg, Winnipeg, MB, R3T 3E2, Canada. E-mail:s.ramanna@uwinnipeg.ca
Abstract: Rough set based flow graphs represent the flow of information for a given data set where branches of these could be constructed as decision rules. However, in the recent years, the concept of flow graphs has been applied to perceptual systems (also called perceptual flow graphs) where they play a vital role in determining the nearness among disjoint sets of perceptual objects. Perceptual flow graphs were first introduced to represent and reason about sufficiently near visual points in images. In this paper, we have given a practical implementation of flow graphs induced by a perceptual system, defined with respect to digital images, to perform Content-Based Image Retrieval (CBIR). Results are generated using the SIMPLicity dataset, and our results are compared with the near-set based tolerance nearness measure (tNM).
Keywords: Content Based Image Retrieval, granular computing, flow graphs, near sets, perceptual system, rough sets
