VISION: Enabling Innovation

atmosphere-plus-heatmap                 subtitle-nlp

Vision is a live and on-demand TV platform developed at Lancaster University, available to both staff and students. With an on-demand library of over 21,000 items and with 40 live channels on offer, Vision presents a compelling alternative to existing broadcaster-backed offerings.

The service also acts as a valuable tool to enable the development of unique functionality. Without the constraints of strict assurance guidelines, researchers are granted the ability to rapidly create and deploy new features. Coupled with the built-in instrumentation, it provides a valuable source of user-generated feedback.

Vision has recently been employed to develop and deploy a number of cutting-edge features. This includes a tool to improve the relevancy of search results by annotating the elements within a video to supplement existing metadata. This is achieved through the parsing of embedded subtitles using established processing techniques. By retrieving recognised phrases and important words, additional information can be displayed to the user during playback, including external links to subjects referenced in the TV show.

Vision has also been used as a tool to provide atmosphere in potentially lonely viewing experiences. This builds upon existing chat-based functionality within Vision to create a sense of community and audience. By recording the text-based chat during each playback, subsequent users can enjoy and interact with the video as if they were watching live. This creates a sense of engagement, regardless of when they are consuming the content. A heat map of the chat is also displayed, allowing viewers to easily seek to potentially interesting or provocative points in a video.

The innovation enabled by Vision continues in the MPAT (Multi-Platform Application Toolkit) project, which aims to significantly reduce the barriers to entry for designing and building device-agnostic media presentations. Part of this effort includes the development of a plug-in repository; examples of which may include the atmospheric capability described above. With this repository, a user can choose to integrate existing plug-ins into their application, further simplifying the authoring process.