← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8952 · 生成模型:GPT-6

随机相位保留频谱结构

理解等级:高级

保留样本的幅度谱、随机化相位再逆变换,可观察仅有频率能量时还剩哪些结构。实值图像要求频谱满足共轭对称,并处理均值与颜色中心化。固定幅度随机相位一般不等于严格高斯抽样;严格高斯模型还需让傅里叶系数按对应方差随机变化。

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

      We can also sample new images by matching the parameter of individual natural images. Figure 27.12 shows four examples. To process color images, we first look for a color space that decorrelates color information for each image. This can be done by applying principal component analysis (PCA) to the three color channels. This results in a 3×3 matrix that will produce three new channels that

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第27章《统计图像模型》(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/c027b.xhtml) - 结论依据:摘录中关于随机相位保留频谱结构的定义、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:We can also sample new images by matching the parameter of individual natural images. Figure 27.12 shows four examples. To process color images, we first look for a color space that decorrelates color information for each image. This can be done by applying principal component analysis (PCA) to the three color channels. This results in a 3×3 matrix that will produce three new channels that