Recognition of shapes by attributed skeletal graphs

摘要: In this paper, we propose a framework to address the problem of generic 2-D shape recognition. The aim is mainly on using potential strength skeleton discrete objects in computer vision and pattern recognition where features are needed for classification. We represent medial axis characteristic points as an attributed skeletal graph model shape. information about object its topology totally embedded them allows comparison different by matching algorithms. experimental results demonstrate correctness detecting computing more regular effective representation perceptual indexing. process, based revised graduated assignment algorithm, has produced encouraging results, showing developed method variety domains. robustness presence scale, reflection rotation transformations prove ability handle noise occlusions.