CONSTRUCTION OF NOISE-REDUCTION FILTER BY USE OF SANDGLASS-TYPE NEURAL-NETWORK

A noise reduction filter composed of a sandglass-type neural network ( Sandglass-type Neural network Noise Reduction Filter: SNNRF) was propo sed in the present paper. Sandglass-type neural network (SNN) has symm etrical layer construction, and consists of the same number of units i n input and output layers and less number of units in a hidden layer. It is known that SNN has the property of processing signals which is e quivalent to KL expansion after learning. We applied the recursive lea st square (RLS) method to learning of SNNRF, so that the SNNRF became able to process on-line noise reduction. This paper showed theoretical ly that SNNRF behaves most optimally when the number of units in the h idden layer is equal to the rank of covariance matrix of signal compon ent included in input signal. Computer experiments confirmed that SNNR F acquired appropriate char acteristics for noise reduction from input signals, and remarkably improved the SN ratio of the signals.