Tracking Faces in Grayscale Video Sequences with "Binary Direction Vectors": Introduction and Evaluation of Binary Direction Vectors as a New Local ... Faces with the "OpenCV CAMSHIFT Tracker" - Softcover

Inhaltsangabe

This diploma thesis introduces a novel, color-independent feature for image analysis.Furthermore, it describes an application prototypeusing this feature for face-tracking and itsextensive evaluation.The main achievement of this thesis is thedevelopment of an alternative, simple and robustbasis for face-tracking solutions and other imageprocessing purposes. This novel method allows theencoding of local, structural features that arerecognizeable in gray-scale images as so-calledBinary Direction Vectors (BDVs). Thisrepresentationof structural information is successfully combinedwith the existing tracking algorithm "OpenCVCAMSHIFT Tracker", to demonstrate the simplehandling of BDVs. The tracking precision of theCAMSHIFT/BDV combination is increased by modifyingthe statistical analysis that is used by thetracking algorithm.The supremacy of the modified version of thetracking algorithm over the original version isproved with an extensive empirical evaluation.Theseevaluation series also demonstrate how trackingsystems can be compared in a precise andscientifically founded way.

Die Inhaltsangabe kann sich auf eine andere Ausgabe dieses Titels beziehen.