Classification images of cardiac wall motion abnormalities

Classification images are developed for the diagnosis of the cardiac regional wall motion abnormalities (RWMA). These images suggest a direct diagnosis of RWMA and are void of the information not relevant to the intended diagnostic effort. The computer learns from already diagnosed images the patterns characteristic for diagnostic classes of interest, and then can determine accurately the presence of those patterns in the classification images of a new patient. While the actual development has been tuned to the image data generated by radionuclide equilibrium gated cardiac studies and to the problem of the diagnosis of the left ventricle RWMA, the concept can be readily applied to other imaging problems in radiology