← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 9065 · 生成模型:GPT-6

运动场为每像素记录位移

理解等级:中级

两帧匹配将第一帧每个像素映射到第二帧的对应位置,位移形成二维向量场。遮挡、新出现区域或外观变化会使对应不存在或不可靠;形式上给出一个向量不代表找到了真实可见对应。

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

      Under this formulation, the task of motion estimation consists of finding, for each pixel in frame 1, the location of the corresponding pixel in frame 2. Just as in the chapter on stereo matching, we need first to define how we will compare pixels to find correspondences.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第46章《运动估计》(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/c046b.xhtml) - 结论依据:摘录中关于运动场为每像素记录位移的定义、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:Under this formulation, the task of motion estimation consists of finding, for each pixel in frame 1, the location of the corresponding pixel in frame 2. Just as in the chapter on stereo matching, we need first to define how we will compare pixels to find correspondences.