Enhancing multimedia protocol performance through knowledge based modeling
Nicola Zingirian, Massimo Maresca, Mauro Migliardi, Pierpaolo Baglietto · 2002
Multimedia transmission protocols take advantage of video compression coders. The core of video sequence compression coders is the motion estimation task. Unfortunately video sequence coders cannot be implemented effectively on desktop workstations because of the high computing requirements of the motion estimation task. It is possible to reduce the computing power required penalizing the compression ratio (i.e. increasing output bandwidth) and/or video quality, but in case of standard software coders running on desktop computers the penalty of compression and quality becomes unacceptable. In order to improve both the compression ratio and processing time we propose to enhance multimedia protocol design through the modeling of the specific application domain. We show the technique in a case study, namely the design of a MPEG coding algorithm targeted to compress video sequences generated by highway surveillance cameras. In addition me show the effectiveness of the technique with experimental results.