direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments


Objective No-reference prediction of saliency changes in the presence of packet losses
Zitatschlüssel Argyropoulos_2013_VPQM
Autor Argyropoulos, Savvas and Garcia, Marie-Neige and List, David and Schleicher, Robert and Raake, Alexander
Buchtitel Int. Workshop on Video Processing and Quality Metrics (VPQM)
Jahr 2013
Zusammenfassung In this paper, a novel no-reference method to predict the impact of packet losses in video sequences on visual attention is proposed. Initially, the results of a subjective eye-tracking experiment are analyzed to determine whether the artefacts caused by network degradations result in abrupt saliency changes. Then, the proposed method employs existing objective visual attention models to compute the saliency maps of the individual frames. Subsequently, the temporal consistency of the saliency maps is investigated and correlated with the error propagation pattern that results from the packet losses to predict visual saliency changes in a rule-based approach. Experimental results on a proprietary database are presented to demonstrate the validity of the proposed method.
Download Bibtex Eintrag

Zusatzinformationen / Extras


Schnellnavigation zur Seite über Nummerneingabe