← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 8951.b · 生成模型:GPT-6

高斯模型表达二阶约束

理解等级:高级

给定均值与正定协方差,在连续分布的相应约束下,高斯分布具有最大微分熵。图像高斯模型由协方差描述像素相关,而不再要求像素独立。书中零均值表达式需先中心化;若协方差奇异,则不能直接套普通逆矩阵和全空间密度。

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 want to write a statistical image model that takes into account the correlation statistics and the power spectrum of natural images (see [439] for a review). If the only constraint we have is the correlation function among image pixels, which mean that among all the possible distributions, the one that has the maximum entropy (thus, imposing the fewest other constraints) is the Gaussian distribution:

    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:摘录中关于高斯模型表达二阶约束的定义、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 want to write a statistical image model that takes into account the correlation statistics and the power spectrum of natural images (see [439] for a review). If the only constraint we have is the correlation function among image pixels, which mean that among all the possible distributions, the one that has the maximum entropy (thus, imposing the fewest other constraints) is the Gaussian distribution: