MULTISPECTRAL CODE EXCITED LINEAR PREDICTION CODING AND ITS APPLICATION IN MAGNETIC-RESONANCE IMAGES

This paper reports a multispectral code excited linear prediction (MCE LP) method for the compression of multispectral images, Different line ar prediction models and adaptation schemes have been compared, The me thod that uses a forward adaptive autoregressive (AR) model has proven to achieve a good compromise between performance, complexity, and rob ustness. This approach is referred to as the MFCELP method, Given a se t of multispectral images, the linear predictive coefficients are upda ted over nonoverlapping three dimensional (3-D) macroblocks. Each macr oblock is further divided into several 3-D micro-blocks, and the best excitation signal for each microblock is determined through an analysi s-by-synthesis procedure. The MFCELP method has been applied to multis pectral magnetic resonance (MR) images, To satisfy the high quality re quirement for medical. images, the error between the original image se t and the synthesized one is further specified using a vector quantize r, This method has been applied to images from 26 clinical MR neuro st udies (20 slices/study, three spectral bands/slice, 256 x 256 pixels/b and, 12 b/pixel). The MFCELP method provides a significant visual impr ovement over the discrete cosine transform (DCT) based Joint Photograp hers Expert Group (JPEG) method, the wavelet transform based embedded zero-tree wavelet (EZW) coding method, and the vector tree (VT) coding method, as well as the multispectral segmented autoregressive moving average (MSARMA) method we developed previously.