发明名称 MOVING IMAGE PROCESSING METHOD
摘要 PROBLEM TO BE SOLVED: To improve the visual recognition by correcting a motion vector obtained at moving image compression into a vector representing a motion between consecutive frames and detecting an image of a frame satisfying a prescribed condition for a degree of dispersion in a frame generated next as a cut of a moving image. SOLUTION: Moving image data are stored in a storage section 104 and a cut detection section 104 detects a cut already and information to be divided into a shot is provided. The user selects how a moving image is to be divided into scenes, by using what image. When a frame is selected and divided into scenes. a similar image retrieval section 107 detects them to be plural similar images. A scene information generating section 106 divides plural shots into scenes by using a frame discriminated to be a similar image as a border. A scene information generating section 106 divides plural shots into scenes, bordering the frame decided to be a similar image. When the same scene is discriminated, it is stored in a storage section 104 and plural frames being similar images are defined by a link generating section 108.
申请公布号 JPH0993588(A) 申请公布日期 1997.04.04
申请号 JP19950250405 申请日期 1995.09.28
申请人 TOSHIBA CORP 发明人 KANEKO TOSHIMITSU;AOKI HISASHI
分类号 H04N5/76;H04N5/92;H04N7/32 主分类号 H04N5/76
代理机构 代理人
主权项
地址