知识卡片

普通读书笔记卡 · 9076 · 生成模型:GPT-6

结构张量秩决定局部可辨识

理解等级:中级

局部A由空间梯度决定:无纹理时近似秩零,单直边时近似秩一,多个独立梯度方向才可能满秩。实际应检查最小特征值及条件数,避免显式求逆放大噪声;满秩也不证明亮度与恒速假设正确。

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

      We will have an equivalent set of equations for each image location. The solution at each location can be computed analytically as u = A–1b where A is the 2 × 2 matrix of equation (48.10). The motion at location x, y will only be uniquely defined if we can compute the inverse of A. Note that the matrix A is a function of the

    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:摘录中关于结构张量秩决定局部可辨识的定义、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:We will have an equivalent set of equations for each image location. The solution at each location can be computed analytically as u = A–1b where A is the 2 × 2 matrix of equation (48.10). The motion at location x, y will only be uniquely defined if we can compute the inverse of A. Note that the matrix A is a function of the