Object Description

The objects of a segmented image are represented by various ways and their features are extracted to enable object recognition. The boundary of an object is represented by chain code, signature and Fourier descriptor. The interior of a region is characterized by their geometrical features, moments and textural features. Typical geometrical features are area, perimeter, compactness and irregularity. Euler number is a topological descriptor. Moments are a sort of combination of geometrical features. Textural features are derived based on the histogram and co-occurrence matrices. Principal component analysis is based on decomposing an image into its components and leads to data reduction using linear algebra techniques.