An Advanced A-V- Player to Support Scalable Personalised Interaction with Multi-Stream Video Content

Abstract

Current Audio-Video (A-V) players are limited to pausing, resuming, selecting and viewing a single video stream of a live broadcast event that is orchestrated by a professional director. The main objective of this research is to investigate how to create a new custom-built interactive A V player that enables viewers to personalise their own orchestrated views of live events from multiple simultaneous camera streams, via interacting with tracked moving objects, being able to zoom in and out of targeted objects, and being able to switch views based upon detected incidents in specific camera views. This involves research and development of a personalisation framework to create and maintain user profiles that are acquired implicitly and explicitly and modelling how this framework supports an evaluation of the effectiveness and usability of personalisation. Personalisation is considered from both an application oriented and a quality supervision oriented perspective within the proposed framework. Personalisation models can be individually or collaboratively linked with specific personalisation usage scenarios. The quality of different personalised interaction in terms of explicit evaluative metrics such as scalability and consistency can be monitored and measured using specific evaluation mechanisms.