http://www.RoboRealm.com/ The newest forum threads. Tue, 30 Nov 1999 00:11:00 EST en-us Copyright: (C) RoboRealm, http://www.RoboRealm.com/ http://www.RoboRealm.com/ There are many different ways to do that. If you are interested in the behinds the scenes way, take a look at the ARTookit <br><br><a href="http://www.hitl.washington.edu/artoolkit/download/" target="_blank">http://www.hitl.washington.edu/<wbr>artoolkit/<wbr>download/<wbr></a><br><br>and you can download their source to have a look at one way of doing it. We use the adaptive threshold to first isolate the pattern and then some shape analysis to extract out the fiducial. Again, there are many ways to do this.<br><br>Yes, you can use ANN to classify the feature once extracted as any classification technique can be used. Typically, because you have a planar rectangular shape that can be unwarped even basic template matching works quite well.<br><br>STeven.<br><br> http://www.roborealm.com/forum/index.php?thread_id=5186 http://www.roborealm.com/forum/index.php?thread_id=5186 Tue, 30 Nov 1999 00:11:00 EST I am curious about the algorithm behind the recognition of Fiducials. What visual features have been extracted, and can we use neural network to do the classification? Thank you very much.<br><br> http://www.roborealm.com/forum/index.php?thread_id=5186 http://www.roborealm.com/forum/index.php?thread_id=5186 Tue, 30 Nov 1999 00:11:00 EST