Two novel active steganalysis schemes which only use one copy of stego image are presented in this paper. We view active steganalysis as blind sources separation (BSS) problem and can be solved with Independent Component Analysis (ICA) algorithm under the assumption that embedded secret message is an independent, identically distributed (i.i.d) random sequence and independent to cover image. With only one copy of stego image we firstly adopt the maximum a posteriori (MAP) estimator to obtain an estimate of cover image. The first scheme views this estimated version as another stego image. The second scheme generates another two signals with the estimated version. All the three signals are as input signals of ICA algorithm. The two schemes overcome the unreasonable constraint of Chandramouli?s method. Experiment results show that the proposed schemes achieve acceptable performance and improve the performance with larger message length.