Log in or register to post comments

Similar Target

December 29, 2014 - 9:40am #1

Hi,

 

VideoPlayback :

we have been using similar targets for two videos.for eg Cover page and inner page.The coverpage has got the DOG IMAGE with a bold title  , the inner page has got the same DOG image without title.But when the target is detected the coverpage is linked to the inner page video and viceversa.If i look in to the SHOW  features ..there are some notable differences.How to increase the accurancy of the target reading???it should ovelay video only if complete features are matched ???not partial...is there a way to do this using code ?

 

Thanks

 

Similar Target

February 9, 2015 - 12:37am #5

No, at this time, the Vuforia API does not expose an API to adjust the "tolerance" (e.g. number of detected features or so...) on the recognition of a target.

Maybe something you could post in the Wish List:

https://developer.vuforia.com/forum/general-discussion/wish-list

 

 

Similar Target

February 6, 2015 - 6:40am #4

Hi allessandro,

 

Thanks for the reply.The problem  is the book is already out in the market.We cant edit the graphics.My only hope is to manage tolrence level.is there any way in the script where i could control the tolrence level of the image recognition.for eg i would like to play the animation only if it recognizes 90 % or 95 of the details or points ?????? 

Similar Target

February 6, 2015 - 6:40am #3

Hi allessandro,

 

Thanks for the reply.The problem  is the book is already out in the market.We cant edit the graphics.My only hope is to manage tolrence level.is there any way in the script where i could control the tolrence level of the image recognition.for eg i would like to play the animation only if it recognizes 90 % or 95 of the details or points ?????? 

Similar Target

January 5, 2015 - 2:19am #2

Vuforia can detect / track a target with some level of tolerance on the total set of features, e.g. even if only a part of the total features are visible in the camera view; this is done to generally ensure good tracking quality and detection capabilities even if the target is not completely (100%) visible (for instance if it is partially occluded by an obstacle or partially out of view). 

When targets are similar and a very significant part of the target is identical (like the "DOG" part in your case), this can lead to the issue you observed. One way to avoid this would be to differentiate the part of the image that contains the DOG  by editing / adding some small details.

 

 

 

Log in or register to post comments