Scholarlink

<p>Point set registration is essential for computer vision, pattern recognition and intelligent robotics. Existing registration methods exhibit limited robustness against non-Gaussian noise, outliers and incomplete point clouds in complex engineering scenarios. This paper proposes a robust rigid point set registration method based on the maximum correntropy criterion (MCC). The method integrates the point-to-plane distance metric into correntropy measurement to construct a novel iterative optimization framework. In each iteration, point correspondences are established via nearest neighbor search, and rigid transformation parameters are optimized under the MCC criterion. An adaptive kernel width updating strategy is adopted to balance global convergence and local alignment accuracy. Experimental results on synthetic and real-world point sets verify the effectiveness and robustness of the proposed method. Relevant codes and data are published to <a href="https://github.com/ygq0000/mcc">https://github.com/ygq0000/mcc</a>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Robust rigid registration via maximum correntropy criterion: a novel approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月29日

查看全文
  • Guiqiang Yang,
  • Jucheng Wang,
  • Ting Gao,
  • Rui Li

摘要

Point set registration is essential for computer vision, pattern recognition and intelligent robotics. Existing registration methods exhibit limited robustness against non-Gaussian noise, outliers and incomplete point clouds in complex engineering scenarios. This paper proposes a robust rigid point set registration method based on the maximum correntropy criterion (MCC). The method integrates the point-to-plane distance metric into correntropy measurement to construct a novel iterative optimization framework. In each iteration, point correspondences are established via nearest neighbor search, and rigid transformation parameters are optimized under the MCC criterion. An adaptive kernel width updating strategy is adopted to balance global convergence and local alignment accuracy. Experimental results on synthetic and real-world point sets verify the effectiveness and robustness of the proposed method. Relevant codes and data are published to https://github.com/ygq0000/mcc.