Various biometric traits, such as fingerprint, face, and iris, have been exclusively used in applications of person verification and identification. Current research in biometrics includes making the verification system more robust and reliable. An existing way of acquiring the biometric sample from the subject needs cooperation from the subject under a controlled environment. However, capturing
... [Show full abstract] a sample in a controlled environment may have constraints of slower identification process, disturbance in person’s movement in public places, and need of extra supervising personnel. One of the main dominant covariates in unconstrained biometrics of the face is the nonuniform illumination across the face sample. In this paper, we have analyzed the effect of nonuniform lighting conditions on the face samples to be recognized. The experiments are performed for different subspace representations of the facial image. The inferences about the effectiveness of feature representation and illumination normalization techniques derived from the experimental observations are useful in developing the face-based unconstrained biometrics.