Ask Your Question
0

Best method to detect facial hair?

asked 2017-04-21 10:24:11 -0500

macc.n gravatar image

Hello everybody, I would like to detect if a person has beard in an image. I extract the face using the Haar Cascade and I have no idea on how to detect the beard. What is the best way to do that?

Thanks for the help.

edit retag flag offensive close merge delete

Comments

I'm also looking for same kind of implementation, "detect if a person has beard in an image". if you can share code of it i think it will help people who is looking to same kind of implementation.

VenkatBommina gravatar imageVenkatBommina ( 2017-10-04 04:52:09 -0500 )edit

@VenkatBommina , please do not post answers here, if you have a question or comment, thank you.

berak gravatar imageberak ( 2017-10-04 04:53:46 -0500 )edit

1 answer

Sort by ยป oldest newest most voted
0

answered 2017-04-21 11:34:45 -0500

The subsequent steps I would follow are

  • After extracting the face, detect facial landmarks, for example by using dlib
  • Then extract features around the chin regio's together with image patches
  • Create a local descriptor, for example an LBP histogram
  • Then do SVM or ANN classification between classes
edit flag offensive delete link more

Comments

1

Thanks for your answer. I'm new with OpenCV, can you explain better the second and the third point?

  • What are the features to extract and how can I extract them?
  • How can i use this histogram? Is it considered as a feature for the training of the classifier?
macc.n gravatar imagemacc.n ( 2017-04-22 06:02:56 -0500 )edit

Well

  • Features are local descriptions of the areas around the facial landmarks. You can use any feature descriptor in OpenCV on top of the found landmark positions.
  • The LBP histogram is just one kind of feature
  • It is indeed a training feature for the classifier
StevenPuttemans gravatar imageStevenPuttemans ( 2017-04-24 05:54:38 -0500 )edit
Login/Signup to Answer

Question Tools

2 followers

Stats

Asked: 2017-04-21 10:24:11 -0500

Seen: 1,679 times

Last updated: Apr 21 '17