Face detection in color images using wavelet packet analysis

We propose a novel scheme for automatic and fast detection of human faces in color images where the number, the location, the orientation and the size of the faces are unknown, under unconstrained scene conditions such as complex background and uncontrolled illumination. First, each frame is segmented using skin chrominance values, providing face area candidates. Then, shape analysis and wavelet packet decomposition are performed on the face area candidates in order to detect human faces. Each face area candidate is described by a subset of band filtered images containing wavelet coefficients. These coefficients characterize the face texture and a set of simple statistical data is extracted in order to form compact and meaningful feature vectors. Then, an efficient and reliable probabilistic metric derived from the Bhattacharrya distance is used in order to classify the face area candidate feature vectors into face or non-face areas.