为什么 Firebase ML Kit 在相同的人脸上每次检测到不同的轮廓值

huangapple go评论75阅读模式
英文:

Why Firebase ML Kit detect different contour values each time for same face

问题

我正在使用Firebase ML Kit开发一个人脸识别应用程序。尽管在谷歌上进行了许多搜索后,我得知Firebase ML Kit并没有提供识别功能。但是我在这里的问题是,我检测到了一个人脸的轮廓,将它们保存到了一个文件中,但是当我再次检测同一个人脸的轮廓时,它们看起来是不同的。为什么会这样?它们应该对于同一个人脸是相同的吧?

英文:

I am developing a face recognition app using firebase ML Kit. Although after doing many searches on google I come to know that Firebase ML kit does not provide functionality of Recognition.
But my question is here, that I detected contours of a face, saved them to a file, but when I again detected contours of same face they appeared different. Why this is so?? They must be same for same face??

答案1

得分: 0

总会存在面部标志位置的差异。例如,由于 API 为每个脸部点提供了 X/Y 坐标,当摄像机与面部之间的角度不同时,坐标也会发生变化。

该 API 实际上是为像面部口罩和滤镜之类的应用而设计的。正如你所说,该 API 不支持面部识别。

英文:

There are always going to be variances in the position of the face landmarks. For example, since the API provides X/Y coordinates for each point in the face, when the angle between the camera and the face is different the coordinates will change as well.

The API is really intended for applications like face masks and filters. Like you said the API does not support face recognition.

huangapple
  • 本文由 发表于 2020年5月5日 18:08:12
  • 转载请务必保留本文链接:https://go.coder-hub.com/61610605.html
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定