Illumination variation is a big problem in face detection which usually requires a costly compensation prior to classification. To avoid this problem we are proposing a method for face detection irrespective of illumination variations. In this context the contribution of the work is twofold. First we introduce illumination invariant Local Structure Features for face detection. For an efficient computation we propose a Modified Census Transform which enhances the original work of Zabih and Wood [10]. Secondly we introduce an efficient face detection classifier for rapid detection to render high performance face detection rate. The Classifier structure is much simpler because we use only single stage classifier than multi-stage approaches, while having similar capabilities. The combination of illumination invariant features together with a simple classifier leads to a realtime processing[12]. Detection results are presented on two commonlyused databases namely BioID set of 1526images and Yale face data base set of 15 people with 11 images for each .We are achieving detection rates of about 99.76% with a very low false positive rate of 0.18%In this paper, we are also proposing a novel hardware architecture of face-detection engine for mobile applications. Here MCT (Modified Census Transform) and Adaboost learning technique as basic algorithms of face-detection engine. The face-detection chip is developed by verifying and implementing through FPGA and ASIC. The developed ASIC chip has advantage in real-time processing, low power consumption, high performance and low cost. So we expect this chip can be easily used in mobile applications.