Abstract
The R-transform is gaining importance in the pattern recognition applications. It is shown in this note that if out of a total of 2” input signal samples, every 2n-t consecutive samples have the same amplitude bounds, then the transform domain can be partitioned into (n-t+l)2 sets such that the amplitude bounds on all the samples in each set are the same. These bounds in the transform domain are related to the bounds in the signal domain. The impact of this on the volume required for the transform storage is discussed.