知识卡片
扩大积分窗会混合边界
更大窗口有机会收集更多梯度方向,缓解局部孔径歧义,却可能跨越对象边界,把不同运动平均在一起。有效区域不保证随窗口单调增加,取决于纹理、归一化与阈值;选择支持域要兼顾条件性和运动一致性。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 光流估计
原书出处 · 1 处
- · 查看原书依据
The regions for which R > 2 will increase by making the apertures larger, which is achieved by using a Gaussian filter, g, with larger σ. However, this will result in smoother estimated flows and the estimated motion will not respect the object boundaries.
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:摘录中关于扩大积分窗会混合边界的定义、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 regions for which R > 2 will increase by making the apertures larger, which is achieved by using a Gaussian filter, g, with larger σ. However, this will result in smoother estimated flows and the estimated motion will not respect the object boundaries.