Shape Recognition Using Segmenting and String Matching

This paper presents an efficient way to represent objects. The image of the object is converted into edge image. Important points curve are identified by dominant point detection method. A line segment every two consecutive important a categorical or non-linear segment. Nonlinear segments fitted as circular arcs. In addition, compactness approximate polygons used feature in shape recognition process. Experimental results show that using this new global has better performance than traditional features such relative distance, length and angle. Overall method effective representing recognizing shapes.