Grid-like fuzzy cluster extraction by extended fuzzy c-lines with local coordinate rotation

Fuzzy clustering has also been applied to object recognition with certain shapes in image processing. This paper proposes a new approach to grid-like cluster extraction, in which gridlike structures are extracted by extended fuzzy c-lines with local coordinate rotation. In the prototype recognition step, four sides of squares are captured in a similar procedure for linear fuzzy clustering after cluster-wise coordination rotation. Cluster memberships are represented by the sum of memberships to sub-clusters that are defined by each side of prototypical squares.