We present a face detection method using spectral histograms and support vector machines. Each image window is represented by its spectral histogram, a feature vector consisting of histograms of filtered images. Using statistical sampling, we show systematically the representation groups face images together; in comparison, commonly used representations often do not exhibit this necessary and desirable property. By using a support vector machine trained on a set of 4500 face and 8000 non-face

more »

... ages, we obtain a robust classifying function for face and non-face patterns. With an effective illumination correction algorithm, our system reliably discriminates face and non-face patterns in images under different kinds of conditions. Our method on two commonly used data sets give the best performance among recent face detection ones. We attribute the high performance to the desirable properties of the spectral histogram representation and good generalization of support vector machines. Several further improvements in computation time and in performance are discussed.