Abstract: A generalized multiresolution likelihood ratio （GMLR） is defined, then the GMI.R test is obtained. The GMLR has the characteristic that can fuse several features which describe different properties, and it can increase distinction between different source

Key words: Generalized multiresolution likelihood ratio （GMLR）, Unsupervised segmentation, Mixture multiscale au toregressive prediction （MMARP） model, Precise