“…The principle is evaluating the coefficients of the original signal associated with a pre-designed basis so that the signal can be represented by few parameters. For example, the wavelet transform [10,13,14,15], the discrete cosine transform (DCT) [16,17], the Hermite transform [18,19], the nonlinear transform [20], the compressed sensing [21,22,23], and the singular value decomposition (SVD) [24,25]. While the wavelet transform is popular, however, the compression performance is largely dependent on the chosen mother wavelet, number of decomposition levels and different optimization schemes [26].…”