Ask Your Question
0

Co-ordinates within the positive images

asked 2015-11-20 03:26:52 -0600

Bhavik gravatar image

updated 2015-11-20 04:03:06 -0600

berak gravatar image

Suppose we want to train a classifier with human face wearing glasses.In this case while creating the .vec file do we need to specifically crop the eyes with glasses or the whole face along with the glasses?

Will be thankful to know the answer

edit retag flag offensive close merge delete

Comments

If you have the faces, just sort them in 2 classes: with glasses / without glasses, (simple way). If this is not giving good results, then you shall go in another way: detecting eyes and classify them in with / without glasses

thdrksdfthmn gravatar imagethdrksdfthmn ( 2015-11-20 04:00:24 -0600 )edit

1 answer

Sort by ยป oldest newest most voted
1

answered 2015-11-20 05:05:26 -0600

It depends on what you actually want!

  • If you want an actual eye+glass detector, then crop all your training data to the eyes with glasses.
  • If you want a face detector, that is able to detect face with and without glasses make a dataset with full faces, both with and without glasses. This will force the system to select features that are common between both classes.
  • If you want to seperate between faces / glassfaces, then cascade classifiers are not the way to go. Then you will need to do some post processing filter after face detection to determine if a glass region is there or not. For that make a second classifier, for example an SVM on a cut out eye or eye+glass region as a binary decision.
edit flag offensive delete link more

Comments

1

ok thank you.

Bhavik gravatar imageBhavik ( 2015-11-20 05:22:23 -0600 )edit

Do accept a solution once you are able to :)

StevenPuttemans gravatar imageStevenPuttemans ( 2015-11-20 05:42:13 -0600 )edit

How to make the second classifier using HOG/SVM to cut out eye or eye+glass region and train the cascade.Although I have gone through http://answers.opencv.org/question/53... link and found some useful links about both SVM and HOG.But I couldn't figure out how to inplement it practically. It is far more comfusing as I couldnot run the hog.cpp in command prompt like the trainclassifier.exe.then how am I going to end up with the results Thanks in advance

Bhavik gravatar imageBhavik ( 2015-11-28 04:40:50 -0600 )edit

Please poor this into a new question, since it does not depend on this one.

StevenPuttemans gravatar imageStevenPuttemans ( 2015-11-30 02:03:00 -0600 )edit

Question Tools

1 follower

Stats

Asked: 2015-11-20 03:26:52 -0600

Seen: 191 times

Last updated: Nov 20 '15