Computer program for fast Karhunen Loeve transform algorithm

The fast KL transform algorithm was applied for data compression of a set of four ERTS multispectral images and its performance was compared with other techniques previously studied on the same image data. The performance criteria used here are mean square error and signal to noise ratio. The results obtained show a superior performance of the fast KL transform coding algorithm on the data set used with respect to the above stated perfomance criteria. A summary of the results is given in Chapter I and details of comparisons and discussion on conclusions are given in Chapter IV.