This paper presents an novel approaches to the detection and recognition of human face and describes an efficient real time colour face recognition system which tracks a subject’s head and then recognizes the person by comparing local texture features of the face to those of know individual. This approach treats face recognition in different colour spaces other than RGB. We have implemented the face detection system using skin colour detection. To avoid the illumination problem and pose constraints in face recognition, two face descriptors are used,i.e., so called colour local Gabor wavelets (CLGWs) and local colour vector binary patterns(LCVBPs). These features are used along with eigen-faces for better recognition. The framework provides the ability to learn to recognize in a unsupervised manne