Scholarlink

<p>This study proposes a new shape descriptor based on the Rotation Profile. The descriptor is a sequence of numbers estimated when the shape is rotated about a fixed point by an angular increment in a range of 180 degrees. For each rotation angle, the number in the sequence is the Jaccard index of the original and rotated shapes. The descriptor is invariant to similarity transformation, which makes it an efficient tool for analyzing various shapes. The study also presents a high-performance implementation of the Rotation Profile construction algorithm. The tests show that the proposed shape descriptor is highly efficient and stable when applied to a range of classification problems, and can be successfully used in computer vision shape analysis applications.</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

Rotation Profile-Based Binary Shape Descriptor

  • 期刊论文

  • 发表日期：

    2025年5月8日

查看全文
  • Oleg Seredin,
  • Nikita Lomov,
  • Daniil Liakhov,
  • Nikita Mityugov,
  • Olesia Kushnir,
  • Andrei Kopylov,
  • Evgenii Semenishchev

摘要

This study proposes a new shape descriptor based on the Rotation Profile. The descriptor is a sequence of numbers estimated when the shape is rotated about a fixed point by an angular increment in a range of 180 degrees. For each rotation angle, the number in the sequence is the Jaccard index of the original and rotated shapes. The descriptor is invariant to similarity transformation, which makes it an efficient tool for analyzing various shapes. The study also presents a high-performance implementation of the Rotation Profile construction algorithm. The tests show that the proposed shape descriptor is highly efficient and stable when applied to a range of classification problems, and can be successfully used in computer vision shape analysis applications.