In this paper the problem of varied illumination and poses is tackled. A person may look like another person from far away at a different angle under varied illumination conditions but, he/she may be a totally different person when viewed under optimum illumination and frontal pose. I have used two pre-processing techniques namely HE (Histogram Equalization) and BPS (Bit Plane Slicing) to process the images before they fed into the post processing system. BPS is used for data compression while HE is used for normalization of the pixels. I have used LDA as the post-processing technique for the purpose of dimensionality reduction and preservation of orthogonality. A comparison of the two techniques is performed and the results are analyzed. The recognition rate, false acceptance rate and the false rejection rate are computed and plotted for three databases namely ORL, FERET and VIT databases. Though bit plane slicing is useful for data compression, histogram equalization showed a higher recognition rate. © 2006–2016. Asian Research Publishing Network (ARPN). All rights reserved.