Log in or register to post comments

object recognition methode

March 26, 2013 - 7:05am #1

dear guys and vuforia developers.
this is my first vuforia forum message and i hope all of your respon on my message.

1. i've been interested in android augmented reality  and gonna build some of application with vuforia, but i was wondering what was the methode used in vuforia for image recognition. 

2. how about markerless augmented reality, should i build database on the application or i can use cloud ? and how to store it.

 

thanks

object recognition methode

March 27, 2013 - 3:51am #4

and for the second question, i would like to use more then 1 objek that appeared from 1 marker, could you help me with that ?

You can enable support for multiple simultaneous image targets by using this code (see code in ImageTargets.cpp):

JNIEXPORT void JNICALL
Java_com_qualcomm_QCARSamples_ImageTargets_ImageTargets_onQCARInitializedNative(JNIEnv *, jobject)
{
    // Register the update callback where we handle the data set swap:
    QCAR::registerCallback(&updateCallback);

    // Comment in to enable tracking of up to 2 targets simultaneously and
    // split the work over multiple frames:
    QCAR::setHint(QCAR::HINT_MAX_SIMULTANEOUS_IMAGE_TARGETS, 2);
}

 

object recognition methode

March 27, 2013 - 2:00am #3

Ok thanks for your respons about image recognition method, i though you could tell me the methode uses for image recognition outside the vuforia dev..

and for the second question, i would like to use more then 1 objek that appeared from 1 marker, could you help me with that ?

object recognition methode

March 26, 2013 - 7:12am #2

Hi,

1. i've been interested in android augmented reality  and gonna build some of application with vuforia, but i was wondering what was the methode used in vuforia for image recognition. 

We cannot disclose the technical details of how the recognition is performed (as Vuforia is not an open source software)

 

2.how about markerless augmented reality, should i build database on the application or i can use cloud ? and how to store it.

The choice between building a device database or using the Cloud will basically depend on the number of targets you want to be able to recognize and track; if it is a few tens of targets (less than 100), you can use a Device Database; if it is about using hundreds or thousands of targets, you should use the Cloud.

Also, if you want to build Targets on-the-fly, you should consider using the User Defined Targets feature (see UserDefinedTargets sample).

I hope this answers your questions.

 

Log in or register to post comments