摘要: Robust gait recognition is a challenging problem, due to the large intra-subject variations and small inter-subject variations. Out of covariate factors like shoe type, carrying condition, elapsed time, it has been demonstrated that clothing most factor for appearance-based recognition. For example, long coat may cover significant amount features make difficult individual In this paper, we proposed random subspace method (RSM) framework clothing-invariant by combining multiple inductive biases classification. Even size training set, can achieve promising performance. Experiments are conducted on OU-ISIR Treadmill dataset B which includes 32 combinations types, average accuracy more than 80%, indicates effectiveness our method.