知识卡片

普通读书笔记卡 · 9029.c · 生成模型:GPT-6

最小特征值判据与Harris有

理解等级:中级

本书以λ_min(M)>阈值筛选特征,这是最小特征值角点判据;经典Harris响应常写detM−k(trM)²。二者共享结构张量,但不是同一评分。λ_min描述单位方向二次变化的下界,实际位移误差还按||Δ||²缩放。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      The smallest eigenvalue, λmin of the matrix, M, indicates the smallest possible change of image intensities under translation of the patch in any direction. Thus, to detect good corners, one can identify all points for which λmin > λc, for some threshold λc. The Harris corner detector has been widely used for identifying features to match across images, although they were later supplanted by the

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第40章《立体视觉》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c040.xhtml) - 结论依据:摘录中关于最小特征值判据与Harris有的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:The smallest eigenvalue, λmin of the matrix, M, indicates the smallest possible change of image intensities under translation of the patch in any direction. Thus, to detect good corners, one can identify all points for which λmin > λc, for some threshold λc. The Harris corner detector has been widely used for identifying features to match across images, although they were later supplanted by the