| 摘要 |
A method and an apparatus for detecting a face on a digital image are provided to detect a face on a digital image accurately by dividing a digital image into a plurality of pieces, and examining and analyzing a color, characteristics, and spatial locations of each piece by stages. A source digital image is successively scaled with various scale factors(S201). Each image generated by scaling is divided into a plurality of pieces(S202). Phased examination is performed to each piece(S203). It is determined whether an image of a piece is a face(S204). If not, it is checked whether the piece candidate is identifiable after printing(S205). It is checked whether a color of the piece is capable of coinciding with a skin color(S206). If not, characteristics of a unique countenance are confirmed and it is checked whether spatial locations of the characteristics are calculated(S207). If confirmation for the characteristics and calculation of the spatial locations of the characteristics are needed, the piece is added to a set of face(S208). |
