Face recognition has many applications in law enforcement, crowd surveillance, security access control and human computer interaction. However, the applications have been, greatly constrained by the limitations of the currently available recognition tools because these tools are either too slow, sensitive to pose, illumination and facial expression, or too expensive, and hence are impractical. Face recognition is difficult and still has a long way to go before it really becomes practical. The aim of this paper is to present how a system for real-time, robust and inexpensive face recognition may be approached. The system contains the following components: Low cost CCTV video cameras or simple digital, cameras to quickly locate faces on 2D face images captured. A constructor of 3D face images using the captured 2D images. Algorithms that accurately match the constructed 3D face images to the 2D face images preciously existing in a gallery in real-time for pose and illumination invariant face recognition.