Linear transformation of positive-valued exponentially-distributed data

We now consider the general linear transformations of spectral or intensity data, which is widely used in signal and image processing and spectral analysis. The most widely-used application of this feature is in spectral and intensity image classification and analysis. An example is the calculation of the band energies as an initial stage in the calculation of MFCC coefficients. Despite the simplicity of the feature transformation, the positive-valued input data makes the problem more challenging and requires a chapter on its own. We use the “energy statistic" (ES) approach of Section 3.2.2. However, in Section 5.4 we consider the case of fixed assumed mean.