Nem vált be? Semmi gond! Nálunk 30 napon belül visszaküldheti
Ajándékutalvánnyal nem nyúlhat mellé. A megajándékozott az ajándékutalványért bármit választhat kínálatunkból.
30 nap a termék visszaküldésére
This diploma thesis introduces a novel, color- independentfeature for image analysis. Furthermore, it describes anapplication prototype using this feature for face-tracking and itsextensive evaluation. The main achievement of this thesis is thedevelopment of an alternative, simple and robust basis forface-tracking solutions and other image processing purposes. Thisnovel method allows the encoding of local, structural features thatare recognizeable in gray-scale images as so-called BinaryDirection Vectors (BDVs). This representation of structuralinformation is successfully combined with the existing trackingalgorithm "OpenCV CAMSHIFT Tracker", to demonstrate the simplehandling of BDVs. The tracking precision of the CAMSHIFT/BDVcombination is increased by modifying the statistical analysis thatis used by the tracking algorithm. The supremacy of the modifiedversion of the tracking algorithm over the original version isproved with an extensive empirical evaluation. These evaluationseries also demonstrate how tracking systems can be compared in aprecise and scientifically founded way.