| 摘要 |
PURPOSE: To easily fetch a composited image by picking up the image of an object regardless of the color of the back by performing image compositing through edge emphasis processing to obtain the edge information of the composited image. CONSTITUTION: An edge emphasizing means 70 inputs image data fetched from a mirror camera 30, performs edge emphasis processing to all picture elements and binarizes them into dark picture elements consisting of the edge of the object and bright picture elements not consisting of the edge. A contour extracting means 80 inputs the dark picture elements from the edge emphasizing means 70, collects the dark picture elements into a linked component picture element as the set of adjacent and linked dark picture elements, numbers them for each linked component picture element later and calculates linked picture element data showing a picture element area surrounded by the centroid coordinates of linked component picture elements and the linked component picture elements. A both-eye recognizing means 90 recognizes specified linked component picture elements corresponding to both eyes, and a feature extracting means 100 calculates the coordinate of a midpoint between the centroids of these specified linked component picture elements and outputs a distance from a central point on the balance of a face to any specified contour part. |
