← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 9074.a · 生成模型:GPT-6

稠密流与轨迹采用不同视角

理解等级:中级

稠密流在固定图像位置记录随时间变化的速度,后续时刻同一坐标可对应不同场景元素。轨迹则跟随同一身份点移动,需要持续维护对应。前者便于逐像素处理,后者便于对象路径分析,不能混作同一种时间序列。

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 quantities u(x, y, t) and v(x, y, t) indicate that a pixel at image coordinates (x, y) at time t moves with velocity (u, v). The problem with this formulation is that we do not have an explicit representation of the trajectory followed by points in the scene over time. As time t changes, the location (x, y) might correspond to different scene elements.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第48章《光流估计》(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/c048.xhtml) - 结论依据:摘录中关于稠密流与轨迹采用不同视角的定义、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 quantities u(x, y, t) and v(x, y, t) indicate that a pixel at image coordinates (x, y) at time t moves with velocity (u, v). The problem with this formulation is that we do not have an explicit representation of the trajectory followed by points in the scene over time. As time t changes, the location (x, y) might correspond to different scene elements.