Summarization of football video content
First Claim
Patent Images
1. A method of processing a video including football comprising:
- (a) a computer identifying a plurality of segments of said video representing an event, wherein said identifying for at least one of said segments includes detecting the start of said event based upon processing of a first single frame of said video, where each of said segments includes a plurality of frames of said video, and wherein said detecting the start of said event is performed without comparing characteristics of any frame of said video to characteristics of frames of model sequences of said event;
(b) verifying that said first single frame is an appropriate start of said event based upon processing of a plurality of additional single frames, each temporally relevant to said first single frame, concurrently with said processing of said first single frame, where detecting the start of said event includes processing a sliding window of frames considered simultaneously, and where the start of said event is identified when at least a threshold percentage of frames in said sliding window are considered candidate hiking frames, said threshold percentage being less than 100%; and
(c) creating a summarization of said video by including said plurality of segments, where said summarization includes fewer frames than said video.
2 Assignments
0 Petitions
Accused Products
Abstract
Summarization of video content including football.
-
Citations
11 Claims
-
1. A method of processing a video including football comprising:
-
(a) a computer identifying a plurality of segments of said video representing an event, wherein said identifying for at least one of said segments includes detecting the start of said event based upon processing of a first single frame of said video, where each of said segments includes a plurality of frames of said video, and wherein said detecting the start of said event is performed without comparing characteristics of any frame of said video to characteristics of frames of model sequences of said event; (b) verifying that said first single frame is an appropriate start of said event based upon processing of a plurality of additional single frames, each temporally relevant to said first single frame, concurrently with said processing of said first single frame, where detecting the start of said event includes processing a sliding window of frames considered simultaneously, and where the start of said event is identified when at least a threshold percentage of frames in said sliding window are considered candidate hiking frames, said threshold percentage being less than 100%; and (c) creating a summarization of said video by including said plurality of segments, where said summarization includes fewer frames than said video. - View Dependent Claims (2, 3, 4, 5, 6)
-
-
7. A method of processing a video including football comprising:
-
(a) a computer identifying a plurality of segments of said video each including a plurality of frames and representing a play initiated by a hiking event, wherein said identifying for at least one of said segments includes individually processing respective frames in a sliding window of a plurality of sequential said frames of said video, said sliding window having a fixed number of frames, and identifying the first frame of a said window of said plurality of sequential frames as the start of said segment when at least a threshold percentage of frames, said threshold less than 100%, in said window are identified as candidate hiking frames; and (b) creating a summarization of said video by including said plurality of segments, where said summarization includes fewer frames than said video. - View Dependent Claims (8, 9, 10)
-
-
11. A method of processing a video including football comprising:
-
(a) a computer identifying a plurality of segments of said video representing an event, wherein said identifying for at least one of said segments includes detecting the start of said event based upon processing of a first single frame of said video, where each of said segments includes a plurality of frames of said video, and wherein said detecting the start of said event is performed without comparing characteristics of any frame of said video to characteristics of frames of model sequences of said event; (b) verifying that said first single frame is an appropriate start of said event based upon individually processing at least one other single frame temporally relevant to said first single frame, wherein said step of verifying includes individually processing respective frames in a sliding window of a plurality of sequential said frames of said video, said sliding window having a fixed number of frames, and identifying the first frame of a said window of said plurality of sequential frames as the start of said segment when at least a threshold percentage of frames, said threshold less than 100%, in said window are identified as candidate frames for the start of said event; and (c) creating a summarization of said video by including said plurality of segments, where said summarization includes fewer frames than said video.
-
Specification