SYNTHETIC-IMAGES BY SUBSPACE TRANSFORMS .I. PRINCIPAL COMPONENTS IMAGES AND RELATED FILTERS

The principal component (PC) approach offers compressions of an image sequence into fewer images and noise suppressing filters. Multiple MR images of the same tomographic slice obtained with different acquisiti on parameters (i.e., with different T-R,T-E, and flip angles), time se quences of images in nuclear medicine, and cardiac ultrasound image se quences are examples of such input image sets. In this paper noise rel ationships of original and linearly transformed image sequences in gen eral, and specifically of original, PC, and PC-filtered images are dis cussed. As the spinoff, it introduces locally weighted PC transforms a nd filters, nonlinear PC's, and a single-image based filter for suppre ssion of noise. Examples illustrate increased perceptibility of anatom ical/functional structures in PC images and PC-filtered images, includ ing extraction of physiological functional information by PC loading c urves. Generally, the more correlated the original images are, the mor e effective is the PC approach.