知识卡片

普通读书笔记卡 · 8922.c · 生成模型:GPT-6

过采样不保证完整

理解等级:中级

输入N维、输出M维的线性表示中,M=N称临界采样,M>N称过采样。能完整恢复输入还要求分析映射具有满列秩;更多系数本身不能排除冗余或信息缺口。“过完备”则同时包含过采样与完整性,不能只按矩阵长宽判定。

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 transform P is said to be critically sampled when M = N. The transform is oversampled when M > N, and undersampled when M < N. The transform P is complete, that is, encoding all image structure, if it is invertible. If critically sampled (i.e., M = N) and the transform is complete, then Q = (PT)–1. If it is overcomplete (oversampled and complete),

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第23章《图像金字塔》(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/c023.xhtml) - 结论依据:摘录中关于过采样不保证完整的定义、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 transform P is said to be critically sampled when M = N. The transform is oversampled when M > N, and undersampled when M < N. The transform P is complete, that is, encoding all image structure, if it is invertible. If critically sampled (i.e., M = N) and the transform is complete, then Q = (PT)–1. If it is overcomplete (oversampled and complete),