Abstract: A computer implemented method of generating at least one highlight video from an input video, comprising, using at least one processor for: identifying a plurality of significant frames of the input video, computing video-level features of the input video, selecting a plurality of subsets of the plurality of significant frames, for each subset, computing a similarity score indicating similarity between visual features of the subset and video-level features of the input video, clustering the input video into a plurality of clusters of sequential frames according to sequential positions within the video based on the similarity scores correlated with the plurality of significant frames, and creating at least one highlight video by selecting a cluster of sequential frames of the input video.
Abstract: A computer implemented method of generating at least one highlight video from an input video, comprising, using at least one processor for: identifying a plurality of significant frames of the input video, computing video-level features of the input video, selecting a plurality of subsets of the plurality of significant frames, for each subset, computing a similarity score indicating similarity between visual features of the subset and video-level features of the input video, clustering the input video into a plurality of clusters of sequential frames according to sequential positions within the video based on the similarity scores correlated with the plurality of significant frames, and creating at least one highlight video by selecting a cluster of sequential frames of the input video.
Abstract: Disclosed herein are methods, and program products for creating automatically a short video clip summarizing highlights of a long video stream, comprising identifying a plurality of topics in a video stream based on analysis of the video stream's content, extracting a plurality of sentences based on analysis of a textual representation of the content, computing a score for each of the sentences indicating a relation of the respective sentence to each of the topics, selecting a plurality of sentence subsets each comprising one or more sentences having a highest score with respect to a receptive one of the topics, selecting a plurality of video sections of the video stream each mapped to the one or more sentences of a respective sentence subset, and creating a video clip by merging the plurality of video sections each relating to one of the plurality of topics.
Abstract: A computer implemented method of selecting significant frames of a compressed video stream based on content difference, comprising, obtaining change information created by an encoder for an encoded video stream constructed of a plurality of encoded frames and indicative of a difference in a visual content between consecutive frames and performing the following for each of the encoded frames to select a plurality of significant frames: (1) analyze the change information to calculate a cumulative difference between the visual content of the respective encoded frame and the visual content of a most recently selected significant frame previously selected from the plurality of encoded frames and (2) select the respective encoded frame as another significant frame in case the cumulative difference exceeds a predefined threshold. Indication for each of the plurality of significant frames is output to one or more analysis systems adapted to analyze the content of the significant frames.