A multi-channel projection haar features method for face detection

The face detection algorithm based on multi-channel map discriminant projection Haar feature is proposed in this paper. The multi-channel map is extracted from the face image, which can reduce the influence of the illumination and the noise in the image. Based on the positive and negative training samples, the enhanced Haar feature are obtained by the linear discriminant projection, which can improve the distinguishing ability of the feature. The response in multi-channel map of the augmented Haar feature in the training samples is calculated based on the two steps mentioned previously. After that, the non-symmetric quick boosting method is used to generate a set of weak classifiers. The weight and threshold of the strong classifier are adjusted by the linear non symmetric classifier. This method not only improves the distinguishing ability of the feature, but also realizes the reasonable division of the non-balanced positive and negative sample space. The experiment indicates that the detection rate using the proposed method increases compared to the original method, which shows more competitiveness than conventional algorithms on FDDB dataset.