Optical techniques, especially the ones related to light scattering, has been seen to capture morphological changes, such as increase in size and density of nuclei in cells. Mueller imaging of the epithelium and basal layer based on polarized elastic scattering in human cervical tissue sections has been used to identify the dysplastic conditions of the human cervix. The effect of dysplasia strongly manifests in the depolarization power and retardance, which differs significantly in normal and dysplastic tissues sections. Principal Component Analysis (PCA) of the depolarization power images derived from polar decomposition of Mueller matrices of 36 patients clearly identifies the tissue region responsible for clear discrimination between the diseased and normal tissues. Significantly, the principal components are found to be sensitive in discriminating normal cervical tissue and the two different stages of dysplastic conditions grade I (GD1) and grade II (GD2) and provide cut-off depolarization values for each of these stages.Though the depolarization values of GD1 are quite random as compared to normal and GD2 states, PCA is able to effectively separate it out by capturing subtle changes in the depolarization values.It is worth noting that in the GD2 stage concentration of cells is high in the epithelial region near the basal layer than the epithelium layer near the surface though this difference between these two regions is not as significant as in GD1. Interestingly, this phenomenon is well reflected in the depolarization values, which PCA uses effectively to segment GD1 and GD2 into different clusters. Retardance values show little variation along the stroma. However, covariance matrix images of dysplastic and normal are able to capture depletion of retardance below the basal layer due to progressive disruption of collagen network in dysplasia.