← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9088.c · 生成模型:GPT-6

检测器快慢比较依赖版本条件

理解等级:中级

书中以早期YOLO说明全图预测与滑窗的效率权衡,其精度排名具有实现与出版时点。不能将某版在某评测上的描述推广为所有YOLO始终较不准确;应按相同输入、指标、硬件与版本比较。

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

      Object localization is an active area of research and there are a number of different formulations that share some elements with the approaches we shared previously. One example is YOLO [403], which makes predictions by looking at the image globally. It is not as accurate as some of the scanning methods but it can be computationally more efficient. There are many other approaches that we

    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:摘录中关于检测器快慢比较依赖版本条件的定义、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:Object localization is an active area of research and there are a number of different formulations that share some elements with the approaches we shared previously. One example is YOLO [403], which makes predictions by looking at the image globally. It is not as accurate as some of the scanning methods but it can be computationally more efficient. There are many other approaches that we