NEURAL MODEL FOR KARHUNEN-LOEVE TRANSFORM WITH APPLICATION TO ADAPTIVE IMAGE COMPRESSION

A neural model approach to perform adaptive calculation of the princip al components (eigenvectors) of the covariance matrix of an input sequ ence is proposed. The algorithm is based on the successive application of the modified Hebbian learning rule proposed by Oja on every new co variance matrix that results after calculating the previous eigenvecto rs. The approach is shown to converge to the next dominant component t hat is linearly independent of all previously determined eigenvectors. The optimal learning rate is calculated by minimising an error functi on of the learning rate along the gradient descent direction. The appr oach is applied to encode grey-level images adaptively, by calculating a limited number of the KLT coefficients that meet a specified perfor mance criterion. The effect of changing the size of the input sequence (number of image subimages), the maximum number of coding coefficient s on the bit-rate values, the compression ratio, the signal-to-noise r atio, and the generalisation capability of the model to encode new ima ges are investigated.