Real-time video summarization by social media tracing and alignment
Publication Date: 2014-Feb-10
The IP.com Prior Art Database
There is an ever-growing need for video summarization. In this disclosure, we propose a novel video summarization approach by leveraging social media data. People discuss in the social media when they are watching real time events. We leverage the discussions in social media for video summarization by align the social media discussion with video segments for: 1) hot segment detection and scoring 2) Generate text summary from social media for hot segment 3) Segment grouping by leverage fans emotions
Page 01 of 3
With the exponential growth of the video chunks, there is an ever-growing need for video summarization. Video summarization helps in efficient storage, quick browsing, and retrieval of large collection of video data without losing important aspects. An important tool to get the main idea and the important scenes without watching the full original video.
There are may existing technologies for video summarization:
- Video summarization by using camera motion
- Video summarization techniques used for meeting Recordings
- Video summarization techniques used for sports video
- Video summarization techniques used for surveillance videos.
- Video summarization techniques used for video-taped presentations.
In our invention, we focus on videos from live broadcast of big events, like sports. Exiting video summarization leverage audio, video, or multimodal. We propose a novel approach to do video summarization by leveraging social media tracing and alignment.
With the emerging of social media, more and more people like to post/share their opinion on the social media when watching a game or live broadcast.
Based on the social media activity (e.g., discussions on a social media by posting/commenting/forwarding) related to the game, we can automatically generate the video summary accordingly by align the social media discussion with the video segments. The steps are described in the diagram below:
1) Peak discussion time monitoring on social media
2) Mapping the...