A method and system to compensate for inaccuracies in cross-well tomography is presented. The method includes obtaining data from at least two receivers in response to transmissions from at least two transmitters. Next, at least one compensated value is derived based on the responses of the receivers to the transmitters. Finally, an inversion is performed based at least in part on the compensated value derived. This method eliminates inaccuracies that can be caused by sensor gain and phase variations in the inversion process. Inversion results with gain and phase compensation produce better imaging results that can better help determine the shape and boundaries of the reservoir.