OpenCV 的 createLBPHFaceRecognizer 中的半径参数是什么？
What is the radius parameter for in OpenCV's createLBPHFaceRecognizer?
文档在 2.4 和 3.0 版本的站点上都截断了中间句子...（link：http://docs.opencv.org/3.0-last-rst/modules/face/doc/facerec_api.html#createlbphfacerecognizer）
radius – The radius used for building the Circular Local Binary
Pattern. The greater the radius, the
我从一篇关于 LBPH 的论文中猜测，半径与每个补丁考虑的区域有关？
我正在尝试为此找出一个合适的值，如果能确切知道不同值对应的是什么就好了。
来自 OpenCV 中的 similar thread：
A small radius of the operator makes the information encoded in the histogram more local. THE GREATER THE RADIUS, the smoother the image but more spatial information you can get.
您可能会发现也很有用 this repository:
Now using an abritrary radius and sample points has two effects. The more sampling points you take, the more patterns you can encode, the more patterns (and probably information) you have, but the higher the computational effort. Instead the larger the radius, the smoother the LBP image, the larger details can be captured, the less discriminative power the description may have (if you don't increase the sampling points at the same time).
文档在 2.4 和 3.0 版本的站点上都截断了中间句子...（link：http://docs.opencv.org/3.0-last-rst/modules/face/doc/facerec_api.html#createlbphfacerecognizer）
radius – The radius used for building the Circular Local Binary Pattern. The greater the radius, the
我从一篇关于 LBPH 的论文中猜测，半径与每个补丁考虑的区域有关？
我正在尝试为此找出一个合适的值，如果能确切知道不同值对应的是什么就好了。
来自 OpenCV 中的 similar thread：
A small radius of the operator makes the information encoded in the histogram more local. THE GREATER THE RADIUS, the smoother the image but more spatial information you can get.
您可能会发现也很有用 this repository:
Now using an abritrary radius and sample points has two effects. The more sampling points you take, the more patterns you can encode, the more patterns (and probably information) you have, but the higher the computational effort. Instead the larger the radius, the smoother the LBP image, the larger details can be captured, the less discriminative power the description may have (if you don't increase the sampling points at the same time).