How to create "trained_classifierNM1.xml"?

asked 2017-11-21 08:48:54 -0600

AlexChan
36 ●1 ●3

updated 2017-11-21 08:49:28 -0600

For text detection, the cpp_file( opencv_contrib/module/text/example/textdetection.cpp) loads "trained_classifierNM1.xml" and "trained_classifierNM2.xml" to create 2 classifer. Does any body know which training set was used to create them?

As far as I know, "trained_classifierNM1.xml" was created by boost with some features, and "trained_classifierNM2.xml" was created by SVM.

The opencv_contrib link: link text The text detection paper link :link text

thanks in advance!

edit retag flag offensive close merge delete

add a comment

1 answer

Sort by » oldest newest most voted

answered 2017-11-21 09:01:51 -0600

berak
32993 ●7 ●81 ●312

updated 2017-11-21 20:06:10 -0600

they were trained with the code from here (this also has (sample) images)

to use it with recent opencv3, you probably need to checkout this pr

edit flag offensive delete link

Comments

thanks! It takes me a long time to find the code, thanks a lot!

AlexChan ( 2017-11-21 19:15:32 -0600 )edit

@AlexChan, no reason of closing down topics that are accepting a correct answer. It is useful to keep them open for possible follow up questions of either you or others.