| 摘要 |
<P>PROBLEM TO BE SOLVED: To reduce a video analysis processing time without deteriorating video analysis accuracy in the parallel processing of video analysis using a physical feature, such as an image data and an acoustic data, provided in a video. <P>SOLUTION: A video analyzer 13 executes in parallel the detection of event information to an analysis target frame in each analysis section defined in an analysis task manager 12. Only when reaching an end frame in the middle of the event information detection, the video analyzer 13 extends the analysis section until the event information is not detected, and continues the detection processing using a duplicated analysis target frame included in the next analysis section. On the occurrence of the extension of the analysis section, an analysis result integrator 14 integrates a plurality of analysis results obtained by parallel execution, using analysis results detected in temporally preceding analysis sections. <P>COPYRIGHT: (C)2013,JPO&INPIT |
