The face detection needs the largest computational cost in processes of automatic face recognition. An estimation of facial position and scale has the trade-off problem between accuracy and efficiency. In this paper, we propose a method that estimates facial position in parallel with facial scale. The method estimates facial position by iterating three processes: estimation using global features, estimation using local features, verification. Facial scale is estimated by using the Scale Conversion in the above iteration process. We demonstrate the advantages of the proposed method through facial detection experiments.
展开▼
