Context Not Content: A Novel Approach to Real-Time User-Generated Video Composition
Denny Stohr, Stefan P. Wilk, Iva Toteva, Wolfgang Effelsberg, Ralf Steinmetz · 2016
Instant sharing of user-generated video recordings has become a widely used service on platforms such as YouNow. Yet, it still poses technical challenges, as mobile upload speed and capacities are limited. One proposed solution to address these issues is video composition. It allows switching between multiple video streams-selecting the best source for a given time-for composing a live video of a better overall quality for viewers. Previous approaches require visual analysis of the video streams, usually limiting the scalability of the system. In contrast, our work allows the stream selection to be realized solely on context information, based on video-and service-quality aspects from sensor and network measurements. The implemented monitoring service for context-aware upload of video streams is evaluated in varying network conditions, with diverse user behavior, including camera shaking and user mobility. We show that a higher efficiency for video upload as well as QoE for viewers can be achieved.