| 摘要 |
PURPOSE: A method for extracting a contour from image data is provided to remove isolated pixels as moving a 3x3 window on image data, and to finally extract a contour for region dividing by connecting cut pixels, then to leave final contour components only necessary for image data transmitting, thereby reducing calculations. CONSTITUTION: A system retrieves as moving a 3x3 window on image data whose contour is extracted(S110), and decides whether central pixels of the window are edge pixels(S120). If so, but if all 8 neighbor pixels are not the edge pixels, the system decides that central pixels are isolated and removes the central pixels(S130,S140). If the central pixels are not isolated and the 8 neighbor pixels have the edge pixels, the system decides whether only one of the neighbor pixels has an edge pixel(S150). If so, the system decides whether a pixel corresponding to a direction '6' and a pixel corresponding to a direction '2' are edge pixels(S160). If so, the system inserts the edge pixels into the central pixels of the window(S170). |
