← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9089.b · 生成模型:GPT-6

IoU衡量相对重叠程度

理解等级:中级

IoU是交集面积除以并集面积,非退化框取值零到一。对两个框一起做同样的平移和正比例缩放,IoU不变;只移动预测框则会改变分数。它衡量框重叠,不证明像素归属或三维实体重建正确。

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

      The IoU is a quantity between 0 and 1, with 0 meaning no overlap and 1 meaning that both bounding boxes are identical. As the IoU is a similarity measure, the loss is defined as:

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第50章《目标识别》(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/c050b.xhtml) - 结论依据:摘录中关于IoU衡量相对重叠程度的定义、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 IoU is a quantity between 0 and 1, with 0 meaning no overlap and 1 meaning that both bounding boxes are identical. As the IoU is a similarity measure, the loss is defined as: