计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (10): 185-192.DOI: 10.3778/j.issn.1002-8331.1901-0448

• 模式识别与人工智能 • 上一篇    下一篇

基于孔洞填充的残缺三维模型检索

陈思琦,刘丽,陈秀秀,张龙,王天时   

  1. 山东师范大学 信息科学与工程学院,济南 250014
  • 出版日期:2020-05-15 发布日期:2020-05-13

Incomplete 3D Model Retrieval Based on Hole Filling

CHEN Siqi, LIU Li, CHEN Xiuxiu, ZHANG Long, WANG Tianshi   

  1. School of Information Science and Engineering, Shandong Normal University, Jinan 250014, China
  • Online:2020-05-15 Published:2020-05-13

摘要:

残缺的三维模型和完整的三维模型在拓扑结构上存在差异,为提高检索方法的鲁棒性,提出基于孔洞填充的残缺三维模型检索方法。通过基于径向基函数的孔洞填充方法对残缺三维模型进行孔洞填充,构造残缺模型对应的完整填充模型;构建填充模型中每个数据点的近邻矩阵,通过求解矩阵获得该点的曲率值;以聚类中心对每个数据点的影响为聚类标准改进[K]-menas算法,对填充模型的数据点进行聚类,引入C_DIS相似度量模型,对不同模型按类别进行匹配检索。实验结果表明,能有效检索出与待检索模型相似的三维模型。

关键词: 残缺三维模型, 曲率, 孔洞填充

Abstract:

In order to improve the robustness of retrieval method, the incomplete 3D model retrieval method based on hole filling is proposed. Firstly, the hole filling method based on radial basis function is used to fill the hole of the incomplete 3D model, and the complete filling model corresponding to the incomplete model is constructed. Then, the nearest neighbor matrix of each data point in the filling model is constructed, and the curvature value of the point is obtained by solving the matrix. Finally, the influence of the clustering center on each data point is used as the clustering standard to modify the [K]-means algorithm. The experimental results show that the 3D model can be retrieved effectively, which is similar to the model to be retrieved.

Key words: incomplete 3D model, curvature, hole filling