In inhomogeneous media, PS converted waves often suffer from severe diodic effects. The traveltime and amplitude of PS converted waves may be different in the forward and reverse shooting directions, giving rise to different stacking velocities of PS converted waves and velocity ratios. These effects, compounded with the asymmetric raypath of PS converted waves, will further increase the difficulties and costs in processing PS converted-wave data. One common method to solve this problem is to separate a data set into two volumes with different shooting directions (e.g., negative or positive offset directions). Different values of the PS converted-wave velocities are used to process the two data sets separately and the two results are combined in the final stage. The problem with this method is that sometimes it is difficult to correlate the data sets and the final combined result may be degraded. In this paper, we propose a method to overcome this problem and apply this method to a 2D data set for improving the PS converted-wave imaging.