知识卡片

普通读书笔记卡 · 9081.e · 生成模型:GPT-6

数据集描述需要版本边界

理解等级:中级

书中列举Middlebury、KITTI、Sintel与Flying Chairs,说明真值采集方式与任务分布的差别。序列数量、最大规模和性能领先是出版时的描述,不能当作当前数据集全貌;复现实验应明确具体版本、划分及有效标签范围。

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 biggest challenge of using a supervised model for motion estimation is that ground truth data is very hard to collect. This is probably one of the main limitations of these approaches. There are some small existing datasets, although this might change in a few years.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第49章《学习运动估计》(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/c049.xhtml) - 结论依据:摘录中关于数据集描述需要版本边界的定义、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 biggest challenge of using a supervised model for motion estimation is that ground truth data is very hard to collect. This is probably one of the main limitations of these approaches. There are some small existing datasets, although this might change in a few years.