知识卡片

普通读书笔记卡 · 8958.b · 生成模型:GPT-6

独立子带统计遗漏联合关系

理解等级:中级

分别匹配每个子带直方图,不能保证边缘在多个尺度对齐、长轮廓连贯或大型元素完整。不同位置和子带可以具有相同边缘分布而联合排列不同。改进表征需显式增加相关统计或上下文约束,而不只是细化同一个直方图。

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 results are often quite good. For images like the triangles of figure 28.5, the algorithm works well. However, we note that correlations of filter responses across scale is not captured in this procedure, and long or large-scale structures are often depicted poorly. Figure 28.9 shows two additional examples with more complex images. To process color images, we first do principal component analysis (PCA) in

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第28章《纹理》(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/c028.xhtml) - 结论依据:摘录中关于独立子带统计遗漏联合关系的定义、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 results are often quite good. For images like the triangles of figure 28.5, the algorithm works well. However, we note that correlations of filter responses across scale is not captured in this procedure, and long or large-scale structures are often depicted poorly. Figure 28.9 shows two additional examples with more complex images. To process color images, we first do principal component analysis (PCA) in