Morphological Shape Description and Shape Recognition Using Geometric Spectrum on Multidimensional Binary Images

A useful morphological shape description tool called geometric spectrum (G-spectrum) for quantifying the geometric features of multidimensional binary images is described. The basis of this tool relies upon the cardinality of a set of non-overlapping segments in an image using morphological operations. G-spectrum is not only used as a shape descriptor for describing the geometric features but also as a tool for shape recognition. In this paper, a shape recognition algorithm using G-spectrum is presented. The test results show that shape recognition using G-spectrum is satisfactory