| 摘要 |
A method of segmenting-summarizing a set of media recordings comprising raw still pictures and raw videos, captured by a media device, such as a smartphone, over a long period of time. The method includes retrieving the set of media recordings from memory and determining characteristics of the media recordings, such as orientation, location, media type, chronology. Weights for the characteristics are applied and represent user's preferences. Weights are used in the generation of segments of a predetermined duration d (e.g. 16 seconds). In a first invention, the weights are used to adjust the amount of video and the timing of the still pictures inside a segment of predetermiend duration d. In a second invention, all segments are representing the set of all corresponding raw media recordings. Segmentation involves the expansion of frames of video or stills to fill up segments of duration d, or the shortening of a video by the deletion of frames to fit into the duration d of a segment. The order of the segments is either chronological or based on a characteristic, for instance media type (first the still pictures, then the videos), or orientation, location, onset of audio. The total duration of the summary is less than the period of time of the raw media capture. A list of the now uniformly segmented media recordings is displayed. |
