← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 8921.b · 生成模型:GPT-6

时空伽博不能唯一辨速

理解等级:中级

匀速平移图像的三维时空频谱位于平面 ωt+vxωx+vyωy=0。单个时空Gabor只选择局部频率团,多个不同速度对应的平面都可能经过这里,因此单个强响应不能唯一决定二维速度。需要组合多个方向与频率的证据,而非只选一个频率峰。

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

      Just as we did with Gaussian derivatives, extending Gabor filter for motion analysis is a direct generalization of the x – y 2D Gabor function to a x – y – t 3D Gabor function. Figure 22.15(a) shows a x – t (cosine and sine) Gabor function in one spatial dimension, and figure 22.15(b) shows a sketch of its Fourier transform. This function is selective

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第22章《滤波器组》(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/c022.xhtml) - 结论依据:摘录中关于时空伽博不能唯一辨速的定义、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:Just as we did with Gaussian derivatives, extending Gabor filter for motion analysis is a direct generalization of the x – y 2D Gabor function to a x – y – t 3D Gabor function. Figure 22.15(a) shows a x – t (cosine and sine) Gabor function in one spatial dimension, and figure 22.15(b) shows a sketch of its Fourier transform. This function is selective