Face Detection Based On HOG-PCA And DPM

Biometric recognition is a reliable and robust personal identity identification method by using individual physiological characteristics. Face detection is one of the most remarkable branches of biometric recognition for its easy to operate and friendly interface. Face detection plays an important role in face-based image analysis and is one of the most important problems in computer vision. Despite the success in the last two decades, the state-of-art face detectors still have problems in dealing with images in the wild for the large appearance variations.Instead of taking appearance variations as black boxes and leaving them to statistical learning algorithms, it propose a structural face model to explicitly represent them. The main research works of the proposed face detection method based on HOG-PCA and DPM as follows:(1) By using a histogram of oriented gradient to describe facial feature, the local variation such as illumination changes will be eliminated. Then we use the principal component analysis to decrease the dimension of HOG feature to increase the detection speed and accuracy.(2) The hierarchical part based structural face model enables part subtype option to describe appearance variations of the local part, and part deformation to capture the deformable variations between different poses and expressions. (3) We use Structured SVM to train the detector. In the process of detection, the input candidate is first fitted by the structural model to infer the part location and part subtype, and the confidence score is then computed based on the fitted configuration to reduce the influence of structure variation.Experiments on the challenging Flickr show that our method has a remarkable performance in detecting faces which have variable poses and expressions.