Partial Shape Classification Using Contour Matching in Distance Transformation
IEEE Transactions on Pattern Analysis and Machine Intelligence
| Hi-index | 0.00 |
In the theory of pattern recognition, one of the basic methods of describing the contour of a plane image of an object is the method of its representation in the form of an ordered set of the most informative points of the contour. These points are commonly called control ones and are generally chosen by way of the analysis of values of the curvature function or its normed estimate, i.e., its weight. In this work, studies are made of the stability of various vector characteristics of a contour to a small change in the weights of control points, to the addition of new control points, etc.