知识卡片

普通读书笔记卡 · 9090.c · 生成模型:GPT-6

平均精度依赖积分协议

理解等级:中级

AP按置信排序形成精确率—召回曲线并汇总其面积,但插值包络、采样点、IoU阈值与类别平均规则会改变数值。应报告具体协议,不能把单类单阈AP与跨类多阈值mAP混用,也不能只看某一阈值精确率替代整曲线。

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 precision-recall curve is non-monotonic. The average precision (AP) summarizes the entire curve with one number. The AP is the area under the precision-recall curve, and it is a number between 0 and 1.

    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:摘录中关于平均精度依赖积分协议的定义、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 precision-recall curve is non-monotonic. The average precision (AP) summarizes the entire curve with one number. The AP is the area under the precision-recall curve, and it is a number between 0 and 1.