知识卡片

普通读书笔记卡 · 8829 · 生成模型:GPT-6

扫描边界分类只是特化启发式

理解等级:中级

在物体不互相遮挡、都落地的简单世界中,从图像上方向下扫描,地到物变化可视为遮挡边,物到地变化可视为接触边。它利用强场景假设提供快速分类;物体叠放或互相遮挡时就可能错,不能直接推广到自然场景。

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

      Figure 2.7: For each vertical line (shown in red), scanning from top to bottom, transitions from ground to figure are occlusion boundaries, and transitions from figure to ground are contact edges. This heuristic will fails when an object occludes another.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第02章《一个简单视觉系统》(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/c002b.xhtml) - 结论依据:摘录中关于扫描边界分类只是特化启发式的定义、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:Figure 2.7: For each vertical line (shown in red), scanning from top to bottom, transitions from ground to figure are occlusion boundaries, and transitions from figure to ground are contact edges. This heuristic will fails when an object occludes another.