Adaptive Streaming & Real-Time Communication

Written by Hao Chen |  25/05/2022 - 10/53

Introduction

Bitrate adaptation algorithms are the primary tool that service providers use to optimize video quality. Despite the significant progress that has been made in the past two decades, the current real-time video communication (RTVC) systems still encounter obstacles to achieving high-quality user experiences (QoE) unfortunately. On one hand, the access network of RTVC end users consists of heterogeneous cellular/WiFi wireless links, all with highly diverse channel characteristics, making the network state highly dynamic over time. While to facilitate live interaction, the luxury of seconds of video buffering is not allowed to be maintained at the receiver side. Instead, only a very limited buffer lasting for a few milliseconds is used to adjust transmission policies against the fluctuating end-to-end network states, which severely weakens the ability to amortize underlying network dynamics.  On the other hand, the content is generated in real-time in RTVC sessions, thus video information (e.g., the size of future video frames) is provided in advance. This further increases the difficulty for RTVC frameworks to guarantee the user QoE. 

Main Contributors:

Hao Chen

Mingliu Sun (PhD 21) 

Bowei Xu (PhD 22)

Yueheng Li (MS 21)

Zicheng Zhang (MS 21)

Qianyuan Zheng (MS 22)

Selected Publications:

  • Yueheng Li, Zicheng Zhang, Hao Chen*, and Zhan Ma, Mamba: Bringing Multi-dimensional ABR to WebRTC, ACM Multimedia, Oct. 2023. [Website]

  • Bowei Xu, Hao Chen*, and Zhan Ma, Karma: Adaptive Video Streaming via Causal Sequence Modeling, ACM Multimedia, Oct. 2023.  [Website]

  • Mingliu Sun, Hao Chen*, and Zhan Ma, Quality-of-Experience Assessment for Ultra-low Latency Live Streaming Videos, IEEE MMSP, Oct. 2023.

  • Qianyuan Zheng, Hao Chen*, and Zhan Ma. Bamboo: Boosting Training Efficiency for Real-Time Video Streaming via Online Grouped Federated Transfer Learning. In Proceedings of the 7th Asia-Pacific Workshop on Networking (APNET '23). Association for Computing Machinery, New York, NY, USA, 215–216.

  • Yueheng Li, Qianyuan Zheng, Zicheng Zhang, Hao Chen*, and Zhan Ma. Improving ABR Performance for Short Video Streaming Using Multi-Agent Reinforcement Learning with Expert Guidance. In Proceedings of the 33rd Workshop on Network and Operating System Support for Digital Audio and Video (NOSSDAV '23), Association for Computing Machinery, New York, NY, USA, 58-64.

  • Chao Zhang, Jiaoyang Yin, Yiling Xu*, Hao Chen*, Xiaozhong Xu, and Shan Liu, OLNC: Online Learning of Network Conditions for Adaptive Video Streaming, 2023 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB), Beijing, China, 2023, pp. 1-6.

  • Zicheng Zhang, Hao Chen*, Xun Cao, and Zhan Ma, Anableps: Adapting Bitrate for Real-Time Communication Using VBR-encoded Video, 2023 IEEE International Conference on Multimedia and Expo (ICME), Brisbane, Australia, 2023, pp. 1685-1690. (Oral Presentation)

  • Hao Chen, Xu Zhang, Yiling Xu, Ju Ren, Jingtao Fan, Zhan Ma, and Wenjun Zhang, T-Gaming: A Cost-Efficient Cloud Gaming System at Scale, IEEE Trans. Parallel and Distributed System, vol. 30, no. 12, pp. 2849-2865, Dec. 2019. [This work has been adopted into a commercialized AnyGame to support thousands of gaming users daily!]



back to top

Email: mazhan@nju.edu.cn Telephone: 025-89681550

2017@Vision Lab, Nanjing University

苏ICP备 10085945 苏公网安备 32011302320427 南信备479号