Name of the inventor: John Ronald Kender
Summarizing Video from Multiple Consecutive Frames: Hierarchical representation of video, which is based on physical locations and camera positions, is described.
The method involves dividing multiple consecutive frames into a sequence of consecutive frames. A mosaic representation is prepared for each sequence of the consecutive frames. Similarities between the mosaic representations are compared. Mosaic representations are clustered into physical settings based upon the similarities between the mosaic representations.
Clustering Video Scenes by Physical Setting:
This provides a technique for summarizing video including temporal and non-temporal representations. It also provides a technique for summarizing video that uses mosaic representation of video in order to cluster shots by physical settings which are efficient and accurate.