Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
537467 | Signal Processing: Image Communication | 2015 | 11 Pages |
•Feature Binding (FB) in the field of pattern recognition is proposed.•FB increases accuracy of face detection.•FB is effective for the obscured face.•FB has a good adaptability for varieties of features.
A face detection and location method based on Feature Binding (FB) is proposed in this paper. The features used for face detection and location are classified and bound into groups. The information of each group is extracted separately during face detection. Through the combination with the constraint relationship, the precise location of the face in the image could be identified by confidence coefficients of all groups. Experimental results show that this proposed method can improve the accuracy rate obviously and has good detection effect on obscured faces. Besides, FB can be good to adapt to varieties of features.