Scholarlink

Max-trees are hierarchical structures for anti-extensive attribute filtering. A node attribute is a function of the connected component of a threshold set represented by the node. Filtering is carried out by removing nodes based on their attributes. In this paper we present a new algorithm for parallel attribute computation and filtering using the subtractive rule and image reconstruction on a GPU. Results show up to \(30.4\times \) speed-up compared to sequential algorithms, and increased numerical stability by several orders of magnitude.

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

Shape Filtering and Max-Tree Attribute Computation on a GPU

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Paul D. Teeninga,
  • Michael H. F. Wilkinson

摘要

Max-trees are hierarchical structures for anti-extensive attribute filtering. A node attribute is a function of the connected component of a threshold set represented by the node. Filtering is carried out by removing nodes based on their attributes. In this paper we present a new algorithm for parallel attribute computation and filtering using the subtractive rule and image reconstruction on a GPU. Results show up to \(30.4\times \) speed-up compared to sequential algorithms, and increased numerical stability by several orders of magnitude.