Shape description from generalized support functions

The generalized support function is considered to be a representation of shape properties of compact connected sets in R2R2. Some interesting properties are studied and several parameters are defined for use in shape description and classification. When these parameters are applied to describe convex figures, they are closely related with the measure of congruent segments of fixed length within the convex figure. Finally, an experimental study is conducted to show the goodness obtained when using the generalized support function in shape classification.

► We propose new functions based on the contour of a set to represent shape properties.
► These functions are based on a generalized support function suitable for general sets.
► We prove interesting properties that show their potentialities to classification.
► The experimental study shows higher rate of effectiveness than tangent angle function.