DocumentCode
162241
Title
Hybrid framework for no-reference video quality indication over LTE networks
Author
Songqing Zhao ; Hong Jiang ; Qiao Cai ; Sherif, Sherif ; Tarraf, Ahmed
Author_Institution
Alcatel-Lucent USA, Murray Hill, NJ, USA
fYear
2014
fDate
9-10 May 2014
Firstpage
1
Lastpage
5
Abstract
This paper proposes an efficient hybrid no-reference video quality indication framework for video transmission over LTE networks. Taking the easily obtained PCAP file from an LTE network node as input, we analyze various header information in the IP packets, extract H.264 payload, and save the payload in a bitstream conforming to the H.264 Annex B format. The H.264 bitstream is further decoded and the quality of the decoded video is measured with no-reference algorithms. Packet level metrics such as packet loss and packet size, bitstream level metrics such as frame error and frame duration, and image level metrics such as blockiness and blur, are obtained using respective efficient and novel algorithms. The metrics from this hybrid framework are crucial for predicting video quality MOS score. Furthermore, the IP packets are mapped to respective video frame, which enables to monitor the effect of network impairments on video quality, troubleshoot network issues, and eventually improve end user QoE.
Keywords
IP networks; Long Term Evolution; quality of experience; video coding; H.264 bitstream; H.264 payload; IP packets; LTE networks; PCAP file; bitstream conforming; bitstream level metrics; blockiness; decoded video; end user QoE; frame duration; frame error; hybrid framework; hybrid no-reference video quality indication framework; image level metrics; no-reference video quality indication; noreference algorithms; packet loss; packet size; troubleshoot network issues; video frame; video quality; video quality MOS score; video transmission; Data mining; Decoding; Image edge detection; Measurement; Payloads; Quality assessment; Video recording; H.264 bitstream; Hybrid No-Reference; LTE; Video Quality; blockiness/blur detection;
fLanguage
English
Publisher
ieee
Conference_Titel
Wireless and Optical Communication Conference (WOCC), 2014 23rd
Conference_Location
Newark, NJ
Type
conf
DOI
10.1109/WOCC.2014.6839951
Filename
6839951
Link To Document