We introduce a novel colour histogram based approach to human face detection in colour images. We reason that different regions of a human face contain different facial features and that the colour distributions of these regions may have certain characteristics that are unique to human faces. By concatenating colour histograms of different regions of a human face, we can form a vector that captures the spatial relationship of these unique regional characteristics, which in turn enables the development of effective face detection schemes. To improve efficiency and detection speed, we use principal component analysis (PCA) to reduce the dimensionality of the histograms and apply skin detection as a pre-processing step to reduce the search space. We use a support vector machine for both skin detection and face detection. We present experimental results to demonstrate the effectiveness of the new method.