← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9087.c · 生成模型:GPT-6

候选提议先控制昂贵计算

理解等级:中级

低成本提议模块筛出可能含对象的区域,再交给较昂贵的分类和框精修。它把计算集中到少量候选,却也给整体召回设上限;第一阶段漏掉的对象一般不能被后续分类恢复。

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 window scanning approach can be slow as the same classifier needs to be applied to tens of thousands of image patches. Selective search makes the process more efficient by proposing an initial set of bounding boxes that are good candidates to contain an object (figure 50.12). This proposal mechanism is performed by a selection mechanism simpler than the object classifier. This approach was motivated

    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:摘录中关于候选提议先控制昂贵计算的定义、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 window scanning approach can be slow as the same classifier needs to be applied to tens of thousands of image patches. Selective search makes the process more efficient by proposing an initial set of bounding boxes that are good candidates to contain an object (figure 50.12). This proposal mechanism is performed by a selection mechanism simpler than the object classifier. This approach was motivated