← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 9061.d · 生成模型:GPT-6

新视图好不等于几何唯一

理解等级:中级

NeRF可同时提供新视图合成和密度结构线索,但有限观测可能被多种场解释。看起来逼真的图像不能单独证明密度是唯一真实几何;若要网格或度量重建,应额外检查表面提取规则、多视图一致性及尺度依据。

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

      In fact, there are two products you can get out of a fit radiance field, one for the graphics community and one for the vision community. The graphics product is a system that can render a scene from novel viewpoints. This works by just applying renderθ (figure 45.11) to new cameras with new viewpoints. The second product is an inferred radiance field, which tells us

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第45章《辐射场》(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/c045.xhtml) - 结论依据:摘录中关于新视图好不等于几何唯一的定义、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:In fact, there are two products you can get out of a fit radiance field, one for the graphics community and one for the vision community. The graphics product is a system that can render a scene from novel viewpoints. This works by just applying renderθ (figure 45.11) to new cameras with new viewpoints. The second product is an inferred radiance field, which tells us