The extraction of filamentary structure from a point cloud is discussed. The filaments are modeled as ridge lines or higher dimensional ridges of an underlying density. We propose two novel algorithms, and provide theoretical guarantees for their convergences. We consider the new algorithms as alternatives to the Subspace Constraint Mean Shift (SCMS) algorithm that do not suffer from a shortcoming of the SCMS that is also revealed in this paper.

翻译：讨论从点云中提取丝状结构的问题。 丝状结构以底密度的脊线或高维脊为模型。 我们提出两种新奇算法,并为它们的趋同提供理论保障。 我们认为,新的算法是子空间约束性中位移算法的替代法,这种算法并不因本文也揭示的SCMS缺陷而受到影响。