Optimization algorithms for detection of dominant points of objects in binary images

In object analyzing studies environmental points of the object provide a lot of useful information and describe the object itself without need for internal points. Dominant points of an object are one of the features that best describe it. The regions of dominant points are those where the most mobility among the boundary points of the objects that are two dimensional images of substances. They are also the characteristic regions that carry the most important information about the object. The subject of this research is the detection of dominant points of objects in digital images. The problem is solved with either 4 new local methods that based on the principle that researching on boundary pixels directly or 2 global methods that an original cost function is defined and iterative resolutions are performed for. In the first part of the study some significant principles and techniques about digital image processing are given. In the second part the proposed local methods are given and the dominant points of a well-known shape in the literature are detected by using the global methods. In addition some new algorithms about generating synthetic polygons that are necessary for applications and testing the methods are given. Finally all the methods are tested by using synthetic polygons which are generated for the purpose that trials can be done as objectively as possible. Simulations and success tests results are given as a cross-table.