← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 9067.a · 生成模型:GPT-6

大块会把运动扩散到背景

理解等级:中级

当块同时包含移动车与均匀道路,辨识强的车纹理可能主导匹配,使附近道路也被赋予车的位移。这是跨运动边界的混合支持问题,不只是随机噪声;增大窗口可能让场更平滑却更不真实。

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

      When using single pixels to represent each location, the matching fails in detecting true correspondences and the estimated motion field is very noisy. Just making the patches 5 × 5 pixels is already capable of detecting many correct matches and the motion field seems to mostly capture the true motion between the two frames. Further increasing the patch size eliminates some of the errors. However,

    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:摘录中关于大块会把运动扩散到背景的定义、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:When using single pixels to represent each location, the matching fails in detecting true correspondences and the estimated motion field is very noisy. Just making the patches 5 × 5 pixels is already capable of detecting many correct matches and the motion field seems to mostly capture the true motion between the two frames. Further increasing the patch size eliminates some of the errors. However,