Biometric face recognition technology has received substantial attention in the past several years due to its potential for a wide variety of applications in both law enforcement and non-law enforcement fields. However, most current face recognition systems are designed for indoor and cooperative-user applications. Moreover, ambient lighting fluctuates greatly between days and among indoor and outdoor environments. Furthermore, illumination is the most significant factor affecting the appearance of faces. Most existing systems, academic and commercial, are compromised in accuracy by changes in environmental illumination. Furthermore, state-of-the-art techniques designed to combat this issue have very low accuracy. This paper attempts to combat the issue by proposing an illumination invariant near infrared face recognition architecture that consists of (1) generating a sequence of directional visibility images using quadrant and circular filters, (2) extracting Local Binary Patterns (LBP) and Histogram of Oriented Gradients (HOG) features, and (3) performing SVM based classification. This technique a) improves the accuracy of the face recognition system, b) works under illumination variations, and c) does not need registration of face information. Furthermore, extensive computer simulations performed on the TUFTS (NIR) database and IIT Delhi NIR Face Database demonstrate that the proposed technique produces 94.52% and 80.41% respectively