您的位置: 首页  >  技术问答  >  两个检测之间的交叉联合

两个检测之间的交叉联合

分类: 技术问答 • 2022-04-20 14:24:43
问题描述：

我正在阅读这篇文章：
Ferrari等人在关联措施部分。我理解，Ferrari et al。尝试通过以下方式获取亲和力：

I was reading through the paper : Ferrari et al. in the "Affinity Measures" section. I understood that Ferrari et al. tries to obtain affinity by :


  1. 位置关联性 - 使用两个检测之间的交叉区域

  2. 外观亲和力 - 使用直方图之间的欧氏距离

  3. KLT点亲和度测量

，我有两个主要问题：

However, I have 2 main problems:


  1. 我不能理解什么实际上意味着交叉联合2检测和如何计算

  2. 我尝试了外观亲和度略有差异。我将RGB检测转换为HSV。将色相和饱和度连接成1个向量，并将其用于与其他检测进行比较。然而，使用这种技术失败了，因为袋子的检测具有比检测到同一个人的头部（具有不同取向）更好的相似性得分。

对上述问题的任何建议或解决方案？非常感谢您的帮助。

Any suggestions or solutions to my problems described above? Thank you and your help is very much appreciated.

答

1）您有两个重叠的边框。您计算框的交集，这是重叠的区域。您计算重叠框的并集，它是整个框的面积减去重叠面积的总和。然后你用交集除以联合。在计算机视觉系统工具箱中有一个功能，名为 bboxOverlapRatio 。

1) You have two overlapping bounding boxes. You compute the intersection of the boxes, which is the area of the overlap. You compute the union of the overlapping boxes, which is the sum of the areas of the entire boxes minus the area of the overlap. Then you divide the intersection by the union. There is a function for that in the Computer Vision System Toolbox called bboxOverlapRatio.

2）一般来说，你不想连接颜色通道。你想要的是一个3D直方图，其中维度为H，S和V.

2) Generally, you don't want to concatenate the color channels. What you want instead, is a 3D histogram, where the dimensions are H, S, and V.

相关推荐