Log in or register to post comments

Marker content/two pass target scan

March 4, 2013 - 11:41am #1

Hi,

I'm trying to understand the best approach for the following requirement.

I need to store some information inside a marker, 32 bit would be great, 20 could be a good starting point. Something like a QR Code, a Micro QR could be an actual option.

One option is to put one million of markers on the cloud but I suspect this is not the best approach and could be really expensive (in CPU time and money).

Another on is to use FrameMarkers but these currently support only 9 bits of information. First question: is it possible to store a little more data using the central part at the API level?

Otherwise I'm considering a "two pass" approach: use a single FrameMarker, detect it, extract the central part and process it to extract the extra data. Looking at the Marker, MarkerResult classes I could not find a simple way to get the bitmap content of the matched area. Is this possible?

Are there other approaches worth considering? I still have to look at the samples, starting with the FrameMarker one, but an early direction could save a lot of time.

 

Thanks in advance

Bye

Lorenzo

 

 

 

 

Marker content/two pass target scan

March 8, 2013 - 10:01am #6

The API doesn't provide methods for creating custom marker encoding schemes. The only supported marker type is a FrameMarker. Most examples of QR code integration have been for Unity - https://developer.vuforia.com/search/apachesolr_search/QR?filters=entity%3Anode+ss_node_type%3Aforum

But you can find open source QR code libraries if you want to see how you can develop your own encoding.

To achieve 99% portable code between iOS and Android, you'll need to use Unity. This will enable you to use C# for both, otherwise you'll be working in different languages.

Marker content/two pass target scan

March 5, 2013 - 2:03am #5

Be sure to post in the forum appropriate to the SDK that you're using (i.e. iOS , Android a/o Unity Extension ).

I'm evaluating the product, target platforms are Android and iOS. I expect to be able to write 99% portable code, correct?

The field of steganography deals with encoding data in images - http://en.wikipedia.org/wiki/Steganography

I'm sorry, my question was not clear. I try again: is there any sample, doc, forum thread with some indications for custom frame marker (for example QR codes or other)?

I mean: should I extend the Marker class? Or start at Trackable level? Looking at the Trackable::TYPE enum looks like the set of Marker types is fixed, but maybe UNKNOWN_TYPE or MARKER could be used in this case. Which method should I override if any to place my custom code?

 

In your previous post you told me that it is possibile to get the pixel data. Can you please point me to an api doc, sample, thread with the details?

 

Marker content/two pass target scan

March 4, 2013 - 1:41pm #4

Be sure to post in the forum appropriate to the SDK that you're using (i.e. iOS , Android a/o Unity Extension ).

The field of steganography deals with encoding data in images - http://en.wikipedia.org/wiki/Steganography

The Vuforia SDK doesn't provide a QR code decoder. You'll need to implement your own.

 

Marker content/two pass target scan

March 4, 2013 - 12:24pm #3

Yes you can obtain the pixel data for an image. Which SDK version are you working with?

2.0.31

 

I expect that it is, but this isn't accommodated by the Vuforia SDK, you'll need to implement your own encoding scheme.

Where should I start to look for this?

 

 

Is there a reason that you don't want to use QR or another existing standard? We've seen these successfully adapted to Vuforia apps.

Mainly for aesthetic reasons. I'd like a marker that stands out, like a logo and I need to store very little data. Anyway for a prototype QR would be perfect: is there any support for these in the SDK? How should I register this kind of trackers?

 

 

 

Marker content/two pass target scan

March 4, 2013 - 11:53am #2

Yes you can obtain the pixel data for an image. Which SDK version are you working with?

First question: is it possible to store a little more data using the central part at the API level?

I expect that it is, but this isn't accommodated by the Vuforia SDK, you'll need to implement your own encoding scheme.

 

Is there a reason that you don't want to use QR or another existing standard? We've seen these successfully adapted to Vuforia apps.

Log in or register to post comments