| 摘要 |
<p>PROBLEM TO BE SOLVED: To provide an image processor for separating a part to be an object without depending upon a manual operation, and to provide an image processing method and program.SOLUTION: The image processor includes a feature point detection part 11 for detecting a feature point from the image of each frame, a corresponding point search part 12 for searching corresponding points between images of each frame, a polygonal shape division part 13 for projecting a plurality of polygonal shapes, an area division part 14 for dividing each image into a plurality of areas on the basis of similarity between pixels with each other in the image of a frame to be an object and the image of a frame in front of or/and at the back of the image of the frame to be the object, an extraction part 15 for extracting each division area corresponding to the position of each polygonal shape as an overlap area, a specification part 16 for specifying a correspondence area corresponding to each overlap area from the image of the frame in front of or/and at the back of the image of the frame to be the object on the basis of the corresponding point, a determination part 17 for determining that a non-rigid body is included in the overlap area, and a separation part 18 for separating a non-rigid body part from the overlap area determined that the non-rigid body is included.</p> |
