知识卡片
颜色聚类成功依赖图像条件
书中a、b聚类能将近无彩色的斑马与带色背景分开,却在颜色交叠的海滩场景中失败。这个对比说明视觉复杂与聚类困难不总一致,关键是选定特征是否分离目标。示例成功不保证对其他光照、背景或相机设置同样有效。
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
- · 知觉分组
原书出处 · 1 处
- · 查看原书依据
Naturally, k-means on ab-values finds segments that correspond to regions of roughly a constant color. This simple approach can be surprisingly effective. notice how it segments a zebra from the background in figure 31.6 (because the black and white stripes have roughly the same ab-values). However, it fails to group more complex patterns that are not defined by color similarity, as in the beach photo
参考来源
- 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第31章《知觉分组》(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/c031.xhtml)
- 结论依据:摘录中关于颜色聚类成功依赖图像条件的定义、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:Naturally, k-means on ab-values finds segments that correspond to regions of roughly a constant color. This simple approach can be surprisingly effective. notice how it segments a zebra from the background in figure 31.6 (because the black and white stripes have roughly the same ab-values). However, it fails to group more complex patterns that are not defined by color similarity, as in the beach photo