知识卡片
带通响应表现尖峰厚尾
自然图像的局部带通或导数响应常聚集在零附近,同时存在少量较大响应。平坦区域贡献小系数,边缘和纹理贡献大系数,因此单个高斯常不能很好拟合。这里的厚尾是相对高斯的经验比较,不应推成每幅图、每个滤波器都满足同一分布。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 统计图像模型
原书出处 · 1 处
- · 查看原书依据
In the 1980s, a number of papers noticed a remarkable property of images: when filtering images with band-pass filters (i.e, image derivatives, Gabor filters, etc.) the output had a non-Gaussian distribution [96, 127]. Figure 27.15 shows the histogram of the input image and the histogram of the output of applying the filters [–1, 1] and [–1, 1]T.
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:摘录中关于带通响应表现尖峰厚尾的定义、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 the 1980s, a number of papers noticed a remarkable property of images: when filtering images with band-pass filters (i.e, image derivatives, Gabor filters, etc.) the output had a non-Gaussian distribution [96, 127]. Figure 27.15 shows the histogram of the input image and the histogram of the output of applying the filters [–1, 1] and [–1, 1]T.