| 摘要 |
<P>PROBLEM TO BE SOLVED: To avoid false detection in a face detection algorithm. <P>SOLUTION: A face video detection device in which a test image from a video sequence is compared with an image property model derived from image properties of a region detected containing a face in a previous image in the video sequence comprises: selecting means for selecting a predetermined proportion of pixels in the region detected containing a face in the previous image which most closely match the image property model derived in respect of that region, to thereby derive a pixel mask; and comparing means for comparing pixels in the test image defined by the pixel mask applied at more than one image position within the test image with the image property model, to thereby detect a face in the test image at a mask position corresponding to the lowest average difference between the image property model and pixels defined by the mask at that mask position. <P>COPYRIGHT: (C)2004,JPO&NCIPI |
