The reconstruction of MRI data assumes a uniform radiofrequency field. However, in practice the radio-frequency field is inhomogeneous and leads to non-biological intensity non-uniformities across an image. This artifact can complicate further automated analysis of the data. In general, an acquisition protocol provides images of the same anatomic region with multiple contrasts representing similar underlying information, but suffering from different intensity non-uniformities. A method is presented for the joint intensity uniformity restoration of two such images. The effect of the intensity distortion on the auto-co-occurrence statistics of each of the two images as well as in their joint-co-occurrence statistics is modeled and used for their restoration with Wiener filtering. Several regularity constrains for the anatomy and for the non-uniformity are also imposed. Moreover, the method considers an inevitable difference between the signal regions of the two images. The joint treatment of the images can improve the accuracy and the efficiency of the restoration as well as decrease the requirements for additional calibration scans. The effectiveness of the method has been demonstrated extensively with both phantom and real brain anatomic data as well as with real DIXON pairs of fat and water abdominal data.