← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9053.b · 生成模型:GPT-6

分辨率决定可见纹理

理解等级:中级

糖罐缩略图看似平滑,高分辨率局部却有可检测的表面不规则。重建前可检查关键点周围裁剪,确认它们来自真实稳定结构,而非噪声与高光。若整套方法依赖局部纹理,完全缺纹理表面会削弱这条管线;这不排除轮廓、投影纹理等其他方法。

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

      Figure 44.5 shows the resulting detected keypoints by SuperPoint on three viewpoints of the sugar bowl. The images shown in figure 44.2 seem to lack texture. Interestingly, when we zoom into the images (which are very high-resolution) we can see that the surface contains many irregularities that are detected by the keypoints classifier (figure 44.6[a]). If the surface of the bowl was smooth and textureless,

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第44章《多视图几何与运动恢复结构》(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/c044.xhtml) - 结论依据:摘录中关于分辨率决定可见纹理的定义、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:Figure 44.5 shows the resulting detected keypoints by SuperPoint on three viewpoints of the sugar bowl. The images shown in figure 44.2 seem to lack texture. Interestingly, when we zoom into the images (which are very high-resolution) we can see that the surface contains many irregularities that are detected by the keypoints classifier (figure 44.6[a]). If the surface of the bowl was smooth and textureless,