知识卡片

普通读书笔记卡 · 9053.c · 生成模型:GPT-6

描述子与检测位置职责不同

理解等级:中级

检测器决定取哪些位置,描述子把该处邻域编码成可跨图比较的向量。书中256维只是所用模型示例,不是所有SIFT、SURF或ORB共同维数。向量接近是候选对应证据,仍需匹配与几何验证,不能直接当同一三维点的证明。

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

      Once we have trained a keypoint detector that finds image regions that are distinct and likely to be stable under geometric transformations, we run the detector in the set of images we want to match. We need now to extract at each location a descriptor that will allow finding the same keypoint across the different images.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第44章《多视图几何与运动恢复结构》(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/c044.xhtml) - 结论依据:摘录中关于描述子与检测位置职责不同的定义、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:Once we have trained a keypoint detector that finds image regions that are distinct and likely to be stable under geometric transformations, we run the detector in the set of images we want to match. We need now to extract at each location a descriptor that will allow finding the same keypoint across the different images.