Graduation Semester and Year

Spring 2025

Language

English

Document Type

Thesis

Degree Name

Master of Science in Computer Science

Department

Computer Science and Engineering

First Advisor

Ming Li

Second Advisor

Chenxi Wang

Third Advisor

Nadra Guizani

Abstract

Video streaming has evolved over the decades to adapt to the increased network quality, user demand, and hardware improvement. Although conventional frameworks have primarily used the state of network or system parameters like buffer size at a given time to decide the quality of the video, they did not consider the user perception on quality due to limitations in the ability to track the user’s watching experience. The introduction of immersive wearable devices opens up a new frontier for video stream- ing systems. Equipped with user-specific sensory data collected from live trackers, this work introduces a new video streaming framework for VR which uses users bio signal along with network to function as a personalized ABR. Bio-signals tracking provides critical insights regarding user’s quality perception in watching videos. By leveraging deep neural networks to analyze these signals, we can effectively infer user experience and tailor the streaming experience to better align with individual percep- tual quality. Real-world and controlled lab testing yield promising results,showcasing the potential as a pioneering personalized video delivery system.

Keywords

Video streaming, DASH, eye tracking, VR, Machine Learning, Adaptive Bitrate

Disciplines

Digital Communications and Networking | Other Computer Engineering

License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Available for download on Saturday, May 08, 2027

Share

COinS