We present a novel solution towards the problem of pose and illumination variation of face detection (FD) and face recognition (FR). In this paper, two advanced method are used to provide pose and illumination invariant FR. The 3D morphable model is implemented to generate 3D face images from our very own training database. This process requires a set of three input face images with varying pose and illumination constraints. The resulting 3D model is then used to train the Support Vector

more »

... (SVM) component-based FR. SVM component-based 3D model has promising results yielding close to 92.6% accuracy when tested on three training face images of each subject under test. Index Terms-Face recognition, 3D model, support vector machine (SVM), component-based recognition.