| 摘要 |
PROBLEM TO BE SOLVED: To precisely retrieve a video image in which a moving object such as an invader appears by recording a featured value expressing the feature of the whole of a varied area even when the varied area detected as the moving object of the invader etc. is divided into a plurality of areas. SOLUTION: When an event detection part 4 specifies a plurality of varied areas which are probably the invader, an extraction candidate area setting part 5 sets the individual varied areas to the extraction candidate area of the featured value, and combines the respective varied areas with each other to set the combined area to the extraction candidate area of the featured value. The featured value of the extraction candidate area set by the extraction candidate area setting part 5 is extracted, and recorded in connection with the video image. COPYRIGHT: (C)2006,JPO&NCIPI
|
