Page 48 - Proceedings of the 2018 ITU Kaleidoscope

2018 ITU Kaleidoscope Academic Conference

Consumer Electronics, vol. 62, no. 4, pp. 380 – 388,
Nov. 2016.
[8] J.Nightingale, P. Salva-Garcia, Jose M. Alcaraz
Calero, and Q.Wang, “5G-QoE: QoE Modelling for
(a) Live video sequesces (original) Ultra-HD Video Streaming in 5G Networks”, IEEE
Trans. on Broadcasting, vol. 62, no. 4, pp. 621 – 634,
April 2018.
[9] M. De Filippo De Grazia, D. Zucchetto , A. Testolin,
A. Zanella , M. Zorzi, and M.Zorzi, “QoE Multi-
Stage Machine Learning for Dynamic Video
(b) Received video sequence (decoded)
Streaming,” IEEE Trans. on Cognitive Comm. and
Networks, vol. 4, no. 1, pp. 146–161, March 2018.
Figure 9 – Some original and decoded frames during live
streaming. [10] M. Gadaleta, F. Chiariotti, M. Rossi, and A. Zanella,
“D-DASH: A Deep Q-learning Framework for
6. CONCLUSION AND FUTURE WORK DASH Video Streaming,” IEEE Trans. on Cognitive
Comm. and Networking, vol. 3, no. 4, pp. 703-718
A HTTP adaptive streaming through 4G wireless network Dec. 2017.
was implemented using Double Sarsa approach of [11] R. S. Sutton and A. G. Barto, “Reinforcement
reinforcement learning. To achieve the better QoE, the Learning: An Introduction,” MIT Press Cambridge,
choices that are made to adjust video quality in a real time Massachusetts, 2012.
video streaming depend on the present condition of the
system and the action to be selected in that state providing [12] H.V. Hasselt, "Double Q-learning," Advances in
maximum reward. The proposed Double Sarsa based Neural Information Processing Systems Conference
learning algorithm utilizing Softmax and e-greedy policy on Neural Information Processing Systems (NIPS
was developed and implemented utilizing ITU-T P.1203.1 2010), Dec. 2010.
model. The results were validated using FR video quality
metrics and proposed method could be recommended in [13] H. V. Hasselt, A. Guez and D. Silver, “Deep
Reinforcement Learning with Double Q-learning”,
standardization of future audio-visual streaming services
over wireless IP network. The system was implemented and DOI: arXiv:1509.06461, Dec. 2015.
tested in one way communication; however, it can [14] M. Dumke, “Double Q(σ) and Q(σ,γ)Unifying
undoubtedly be employed to facilitate two-way video Reinforcement Learning Control Algorithms”, DOI:
communication. arXiv:1711.01569v1 5, Nov. 2017.
[15] M.Ganger, E. Duryea, and W.Hu, “Double Sarsa and
7. REFERENCES
Double Expected Sarsa with Shallow and Deep
[1] Cisco Visual Networking Index: Forecast and Learning”, Journal of Data Analysis and Information
Methodology, 2016–2021, Updated on June 6, 2017 Processing, vol. 4, no.4, pp. 159-176, Oct. 2016.
[2] Ericsson Mobility Report, Interim Update, Nov. [16] ITU-T Recommendation P.1203, “Parametric
2017. www.ericsson.com/mobility-report bitstream-based quality assessment of progressive
download and adaptive audiovisual streaming
[3] Global Internet Report, “Paths to Our Digital
Future,” Internet Society, 2017. services over reliable transport”, 2016.
[17] L.Yu, T.Tillo, J. Xiao, “QoE-Drien Dynamic
[4] ISO/IEC JTC 1/SC 29/WG 11, “Coding of Moving
Pictures and Audio” ISO/IEC JTC 1/SC 29/WG 11 Adaptive video Streaming Strategy with Future
N12258, Dec. 2011. Information,” IEEE Trans. on Broadcasting Society,
vol. 63, no. 3, pp. 523 – 534, Sep. 2017.
[5] W. Huang, Y. Zhou, X. Xie, Di Wu, M. Chen, and E.
Ngai, “Buffer State is Enough: Simplifying the [18] ITU-T Recommendation J.247, “Objective
Design of QoE-Aware HTTP Adaptive Video Perceptual Multimedia Video Quality Measurement
Streaming,” IEEE Trans. on Broadcasting, vol. 64, in the Presence of a Full Reference,” Aug. 2008.
no. 2, pp. 590 – 601, June 2018. [19] ISO/IEC JTC1/SC29/WG11 and ITU-T SG16 Q.6
JVT-AB031, “New Video Quality Metrics in the
[6] G. Tian and Y.Liu, “Towards Agile and Smooth
Video Adaptation in HTTP Adaptive Streaming,” H.264 Reference Software,” July 2008.
IEEE/ACM Trans. on Networking, vol. 24, no. 4, pp. [20] “DirectShow (Windows)”.
2386–2399, Aug. 2016. https://docs.microsoft.com/en-
gb/windows/desktop/DirectShow/directshow
[7] V.Martín, J. Cabrera, and N.García, “Design,
optimization and evaluation of a Q-Learning HTTP [21] http://jnetpcap.sourceforge.net/docs/jnetpcap-1.0-javadoc/
Adaptive Streaming Client,” IEEE Trans. on [22] https://www.jio.com/shop/en-in/jiofi/p/491193576

– 32 –