Multidimensional Tensor Product and FFT

Multidimensional signal processing is a child of the computer revolution. Several fields of application such as weather, crystallography, sonar/radar, medical imaging, and seismology require multidimensional data arrays for complete and faithful modeling. The processing of such multidimensional data, one dimension at a time, introduces an error-prone and time-consuming reconstruction stage. Redundancies introduced by time-frequency, time-scale, and space-frequency analysis of speech and vision data naturally lead to multidimensional computations.