摘要 |
Prepare digital data of a value of each pixel within an object image region including a region of a human face. Then, successively, in the object image region, extract position of a Between-the-Eyes candidate point through a filtering process with a Between-the-Eyes detecting filter in which six rectangles Si (1<=i<=6) are connected. Further, extracting a portion of the object image in a prescribed size which has the extracted position of the Between-the-Eyes candidate point at a center, and select a true candidate point from said Between-the-Eyes candidate points in accordance with a pattern discriminating process.
|