Homocentric Polar-Radius Moment for Shape Classification

The description of the object shape is an important subject in the area of image processing and pattern recognition. In this paper, a new approach to contour recognition is proposed. This approach is based on the homocentric polar-radius moment to extract feature vectors of boundaries. Firstly, the definition of homocentric polar-radius moment is introduced. Secondly, the invariance properties of translation, rotation and scaling are proved. In the classification experiments, the recognition results show that our method is more robust and precise than improved moment invariants