知识卡片

普通读书笔记卡 · 8898.g · 生成模型:GPT-6

本征图像恢复物理原因

理解等级:中级

本征图像分解试图从观测图像中恢复反射率、阴影、深度、法线、遮挡或反射等场景属性。Retinex 的照明与反射率分解是其中一个受强先验限制的例子。把图像拆成两个可相乘的数组只是数值分解,赋予物理意义还需要模型假设与可检验约束。

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

      Decomposing an image into different physical causes is known as intrinsic images decomposition. The intrinsic image decomposition, proposed by Barrow and Tenenbaum [39], aims to recover intrinsic scene characteristics from images, such as occlusions, depth, surface normals, shading, reflectance, reflections, and so on.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第18章《图像导数》(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/c018.xhtml) - 结论依据:摘录中关于本征图像恢复物理原因的定义、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:Decomposing an image into different physical causes is known as intrinsic images decomposition. The intrinsic image decomposition, proposed by Barrow and Tenenbaum [39], aims to recover intrinsic scene characteristics from images, such as occlusions, depth, surface normals, shading, reflectance, reflections, and so on.