. Skin colour detection is an often used cue in human motion tracking, especially in face tracking. Skin colour detection is orientation invariant and fast to process. In this paper skin colour is modelled based on a reectance model of the skin, knowledge about the camera parameters, and the spectrum of the light source. In particular, the location of the skin colour area in the chromaticity plane is estimated for dierent light sources, given known camera characteristics. The model is empirically validated. It has application in adaptive segmentation of skin colour and may have application in the estimation of the colour temperature in camera images containing skin colour.

In this article, we bring forward a novel model-based region segmentation algorithm in colour images and use it to segment the skin-colour area from the background area adaptively. First, a quantitative colour space selection scheme is used to select a suitable colour representation manner. A two-fuzzy-set colour distribution model is adopted to depict the object area colour distribution in the selected colour space. Segmentation is performed by use of the α-cuts of the fuzzy sets. The value of α is adaptively decided to different images. This method is successfully used to segment skin colour area in arbitrary colour images with complex background and varying lighting conditions. 1

. We present a novel pre-filtering technique that identifies probable frontal illuminated face regions in colour images regardless of translation, orientation, and scale. The face candidate regions are normalised and provide the basis for face verification using published face detection algorithms. The technique focuses on a fast search strategy to locate potential eye-pairs in an image or video frame. The eye-pair candidates indicate areas that may contain faces. Scale and orientation is inferred from the eye-pairs, and a neural network is used to con rm the normalised face candidates. 1 Introduction Detecting the presence of human faces can provide important cues for many image and video analysis tasks[1]. We are interested in enhancing multimedia tasks, in particular content-based video browsing, retrieval and navigation but face detection and location is also used as a pre-requiste for face analysis tasks such as recognition and expression interpretation. It is non-trivial i...