| 摘要 |
When identifying an object in an image, very often redundant identifications of the object may occur. The present invention relates to methods of image analysis for resolving such redundant identifications of an object. More specifically, the present invention relates to object identification schemes which obtain multiple representations of the same object. In addition, such identification schemes may identify void, or donut-shaped objects when a single search, or multiple searches, of the image are performed. Further, the present invention is useful in resolving several distinct objects where other identification schemes employing either single or multiple searching have identified these distinct objects as a single clumped object. The present invention may be used, for instance, to determine whether a smaller object is contained within another, larger object. |
