Face_recognition: dlib方法没有mtcnn效果好吧

[Allenstin]

• face_recognition version:
• Python version:
• Operating System:

Description

Describe what you were trying to get done.
Tell us what happened, what went wrong, and what you expected to happen.
IMPORTANT: If your issue is related to a specific picture, include it so others can reproduce the issue.

What I Did

Paste the command(s) you ran and the output.
If there was a crash, please include the traceback here.

[Allenstin]

如题.我记得facenet那个文档里面说使用dlib方法的泛化性并不好,不如mtcnn方法好.但为啥这个方法的星星最多呢.值得思索.

[kelvinjjwong]

Obviously the author does not know Chinese.
Allenstin asked why not use mtcnn instead, as he heard that dlib has weaker extensibility than mtcnn possibly.
Maybe Allenstin would ask how to integrate mtcnn as well?

[kelvinjjwong]

But I personally would like to use dlib as I saw some posts clarified the comparison btw dlib and mtcnn on performance in cpu mode and gpu mode.
And I really like this wrapper that makes a lot of things much easier.

[Allenstin]

But I personally would like to use dlib as I saw some posts clarified the comparison btw dlib and mtcnn on performance in cpu mode and gpu mode.
And I really like this wrapper that makes a lot of things much easier.

我对dlib并不清楚。mtcnn已经有封装好的代码了，你直接在github里面搜facenet，那个作者的人脸检测就是用mtcnn来做的。我知道作者可能不懂中文，我写中文就没打算让那些不懂中文的来看，我觉得在国外网站说中文，是一件愉快的事情。谢谢

[hsm4703]

i want to know dlib 68 landmark points rather than like you say github facenet mtcnn 5points which Recognition performance good ，Reasonable 68 points can detect more fearture Relatively mtcnn only have 5 points