Objective video quality assessment: a neural-network approach

Paolo Gastaldo, Stefano Rovetta, and Rodolfo Zunino

DIBE - Dept. Biophysical and Electronic Engineering - University of Genoa
Via all'Opera Pia 11a - 16145 Genova - Italy
dibe.unige.it



In: Proceedings of the International Conference of Media Futures 2001
Florence, Italy - May 8-9, 2001



A B S T R A C T

The paper presents a tool that uses feed-forward neural-network
technologies for the objective quality assessment of MPEG-2 video.
The system operates on a frame-by-frame basis and yields continuous
output, providing a real-time monitoring tool for displayed video
quality. Features used for the estimation have been designed according
to their assessed relevance to perceived quality. 


Keywords: objective quality assessment, neural network, MPEG-2 video.