| 摘要 |
In a system to detect eye position in a facial image comprising a plurality of pixels each represented by a grey level value, a window generator generates windows of the facial image and a histogram extractor creates a histogram for each window of number of pixels vs grey level value. When the number of peaks in the histogram equals 3, the sum of the areas of the peaks having the maximum and minimum grey level is calculated and co-ordinate information corresponding thereto is stored. The eye position is determined from the co-ordinate information corresponding to the window having the largest calculated area. |
