TY - JOUR T1 - On Standardizing the MR Image Intensity Scale JF - RADIOLOGY Y1 - 1998 AB - PURPOSE: MR image intensities have varying ranges and meaningeven for the same protocol (P) and body region (D). This causes many difficulties in image display and analysis. This exhibit describes a method of standardizing the intensity scale, so that for the same P and D, similar intensities will have similar meaning. MATERIALS AND METHODS: In the TRAINING phase (done only once for a given P and D), the parameters of the standardizing transformation are "learnt" from an image set. In the MAPPING phase, done for each MR study, these parameters are utilized to determine the mapping needed to deform its histogram into the standardized histogram. The method was tested quantitatively on 90 brain FSE T2, PD and T1 studies of MS patients and qualitatively on an additional 15 SE PD, T1 and SPGR studies of the brain and foot. RESULTS: As measured by mean squared difference, standardized images have statistically significantly (p<0.01) more consistent range and meaning than those without. Fixed windows that do not require per study adjustment can be established for the standardized images. CONCLUSIONS: Standardizing MR intensity scales to overcome the difficulties due to widely varying intensity meaning is feasible by protocol and body region. This can be implemented in a PACS via DICOM value of interest look up tables. VL - 209 SN - 0033-8419 IS - SUPPL P N1 - 84th Scientific Assembly and Annual Meeting of the RadiologicalSociety of North America (RSNA) Chicago, IL, USA, 1998.11.29-1998.12.04. JO - RADIOLOGY ER -