Hybrid Images for Personalized Media Streaming Optimization

Doh-Suk Kim, Scott Daly, Ludovic Malfait, Neel Chaudhari, Jeffrey Riedmiller

Efficient delivery of media while maintaining or improving the quality of experience (QoE) poses a constant challenge. On the media playback side, various factors, such as device characteristics, viewing conditions, and individual visual acuity, affect QoE. This paper proposes a novel method to estimate viewers' visible frequency ranges to optimize media delivery tailored to individual users and playback contexts. Our approach utilizes hybrid images created by combining high and low-frequency components of two different images. By measuring user perception in hybrid image tasks, we can estimate the minimum video resolution required to maintain quality, reducing unnecessary data transmission without sacrificing user experience. Our experimental results demonstrate a strong correlation between the proposed method and traditional approaches while offering the advantages of lower complexity of cognitive process and user effort. This technique paves the way for personalized and adaptive media delivery in various viewing conditions and on diverse devices.

Print ISSN
Electronic ISSN
2160-2492
Published
2024-07
Content type
Original Research
Keywords
qoe (quality of experience), hybrid image, playback-side context, image resolution
DOI
10.5594/JMI.2024/DPLO3149