Evaluation Criteria Of 3D Mesh Segmentation Under Misclassified Error And Ultimate Measurement Accuracy

Recently, the study of 3D mesh segmentation has been achieved large achievement. But it has not appeared general segmentation algorithms or basic theory yet. The existing segmentation algorithms all face the specific application requirements or specific 3D mesh model proposes and they have lower universality. In addition, the existing segmentation algorithms how to choose the most effective segmentation methods is still an unsolved problem in dealing with specific application of specific grid models. Then 3D mesh segmentation evaluation methods are put forward. It includes two sections: first, analysis in the same segmentation multiple segmentation algorithms is the performance of the mesh and compare inferiority of each algorithm in order to help users in different segmentation applications to select the most suitable 3D mesh segmentation algorithms. Second, it uses the segmentation evaluation results to improve the performance of the existing algorithms and the segmentation quality and guide the new algorithm of research.The study mainly includes three aspects as below:(1) The paper summarizes the 3D mesh segmentation research background and content and reviews the development of 3D mesh. It summarizes the existing 3D mesh segmentation evaluation methods.(2) Through 3D mesh segmentation evaluation method for the classification comparison, we understand it developing in recent years. 3D mesh segmentation evaluation method can guide users to choose the most suitable in segmentation 3D mesh segmentation algorithms and the research of the new algorithms. So researchers become to notice research of computer graphics in the field. The paper proposes misclassified error and the measurement precision of the ultimate 3D mesh segmentation evaluation method. The two methods are applied to evaluation by 3D mesh segmentation evaluation from image segmentation.(3) The paper presents two methods which use 3D mesh library of the Princeton University through the judgment segmentation library adopting the automatic segmentation algorithm. They get actual mesh segmented regions and benchmarks area matching degree to evaluation segmentation algorithm of seven currently popular consumers. The automatic segmentation algorithm was quantitatively experiment evaluation.