A reliable open-source face tagging training and recognition program

Learn from: http://www.cvchina.info/2011/04/15/facel/

He turned to awesome Yang. Please slam him.

In the past few years, we have used opencv for face detection,CodeIt can be found everywhere on the Internet. There are actually two lines of code. However, if you want to find complex andReliableThe computer vision Code related to the face is very rare. Complex: At least it's not like AdaBoost + camshfit. It's reliable:AlgorithmDo not use more fancy, work (it is difficult to do this in many cases), and the code writing is regular and there are comments in the document. I know one or two of these open-source projects. It's strange that their exposure is low. Let's share them with you.

This is facel. On the project homepage, author David S. bolme
And
J. Ross Beveridge is from Colorado State University. Facel is a video capture, face tagging, training, and classification based on the opencv Python interface.One StopSoftware. From official videos, the results are very reliable! But there are only a few hundred views... I am really wondering, there are more than 8000 page views in my weak face recognition demo at Jiaotong University... Bytes

A label can be a person's identity or expression, or you can use a label with or without sunglasses. A classifier is a tag. Facel uses the python interface of libsvm, which is directly integrated for training and online in real time.

To achieve the above functions, there are ready-made algorithms. It seems that facel's credit is to save everything, but this is enough to save a lot of time for Computer Vision practitioners. However, facel also implements an eye tracker named ASEF, which is very reliable and accurate, and is a convolution. The algorithm comes from the author's paper: average on cvpr in.
Of synthetic exact extends lters. I think this is a very useful thing. I just port it into a C code and it is not vague to run it on the iPhone.

The project is developed on Mac, but both Windows and Linux can run. In the end, facel is awesome!