Multiple Classifiers Approach for Computational Efficiency in Multi-scale Search Based Face Detection

Abstract

The multi-scale search based face detection is essential to use a window scanning technique where the window is scanned pixel-by-pixel to search for faces in various positions and scales within an image. Therefore, detection of faces requires high computation cost which prevents from being used in real time applications. In this paper, we present face detection approach by using multiple classifiers for reducing the search space and improving detection accuracy. We design three face classifiers which take different feature representation of local image: gradient, texture, and pixel intensity features. The designed three face classifiers are trained by error back propagation algorithm. The computational efficiency is achieved by coarse-to-fine classification approach. A coarse location of a face is first classified by the gradient feature based face classifier where the window is scanned in large moving steps. From the coarse location of a face, the fine classification is performed to identify the local image as a face where the window is finely scanned. In fine classification, the output of each face classifier is combined and then used for a reliable judgment on the existence of face. Experimental results demonstrate that our proposed method can significantly reduce the number of scans compared to the exhaustive full scanning technique and provides the high detection rate.