This research study is focused on investigating the problem of recognition of human identity from its walking pattern in presence of view angle and clothing condition as covariate. The problem of gait recognition is challenging due to fact that gait has spatio-temporal phenomena with very high dimensional tensorial data distribution, large amount of redundancy, complex pattern distribution and very large variability of appearance within the same class of subject. The extraction of discriminative features in the presence of covariates for robust human gait recognition is a challenging task. The study presents processing aspects of gait recognition system through systematic framework of pre-processing, gait representation, feature extraction, dimensionality reduction and classification for human identity recognition. It has contributed to understanding, interpretation and development of effective gait representation and multilinear subspace learning algorithm. The key contribution, understanding and interpretation are summarized here and direction for future work is provided.